Coffee Quality: Cultivars, Blends, Processing, and Storage Impact

Journal of Food Quality

Lead Guest Editor: Gabriel H. H. De Oliveira Guest Editors: Ana P. L. R. De Oliveira, Fernando M. Botelho, Pedro C. Treto, and Silvia C. C. Botelho Coffee Quality: Cultivars, Blends, Processing, and Storage Impact Journal of Food Quality

Coffee Quality: Cultivars, Blends, Processing, and Storage Impact

This is a special issue published in “Journal of Food Quality.” All articles are open access articles distributed under the Creative Com- mons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Editorial Board

Encarna Aguayo, Spain Vita Di Stefano, Italy Antonio Piga, Italy Riccarda Antiochia, Italy Hüseyin Erten, Turkey Witoon Prinyawiwatkul, USA Jorge Barros-Velázquez, Spain Susana Fiszman, Spain Eduardo Puértolas, Spain José A. Beltrán, Spain Andrea Galimberti, Italy Anet Režek Jambrak, Croatia Luca Campone, Italy Efstathios Giaouris, Greece Juan E. Rivera, Mexico Á. A. Carbonell-Barrachina, Spain Vicente M. Gómez-López, Spain Flora V. Romeo, Italy Marina Carcea, Italy Elena González-Fandos, Spain Jordi Rovira, Spain Márcio Carocho, Portugal Alejandro Hernández, Spain Antonio J. Signes-Pastor, USA Maria Rosaria Corbo, Italy Francisca Hernández, Spain Amy Simonne, USA Daniel Cozzolino, Australia Vera Lavelli, Italy Barbara Speranza, Italy Egidio De Benedetto, Italy Jesús Lozano, Spain Antoni Szumny, Poland Alessandra Del Caro, Italy Sara Panseri, Italy Giuseppe Zeppa, Italy Antimo Di Maro, Italy Luis Patarata, Portugal Dimitrios I. Zeugolis, Ireland Rossella Di Monaco, Italy María B. Pérez-Gago, Spain Teresa Zotta, Italy Contents

Coffee Quality: Cultivars, Blends, Processing, and Storage Impact Gabriel Henrique Horta de Oliveira , Ana Paula Lelis Rodrigues de Oliveira , Fernando Mendes Botelho, Pedro Casanova Treto , and Silvia de Carvalho Campos Botelho Editorial (1 page), Article ID 9805635, Volume 2018 (2018)

Boron, Copper, and Zinc Affect the Productivity, Cup Quality, and Chemical Compounds in Coffee Beans Junia Maria Clemente, Herminia Emilia Prieto Martinez , Adriene Woods Pedrosa , Yonara Poltronieri Neves, Paulo Roberto Cecon, and John Lonfover Jifon Research Article (14 pages), Article ID 7960231, Volume 2018 (2018)

Quality of Commercial Coffees: Heavy Metal and Ash Contents Mariana Teixeira Pigozzi, Flávia Regina Passos , and Fabrícia Queiroz Mendes Research Article (7 pages), Article ID 5908463, Volume 2018 (2018)

Effect of Yeast Fermentation of Green Coffee Beans on Antioxidant Activity and Consumer Acceptability Han Sub Kwak ,YoonhwaJeong ,andMisookKim Research Article (8 pages), Article ID 5967130, Volume 2018 (2018)

Shade Trees Spatial Distribution and Its Effect on Grains and Beverage Quality of Shaded Coffee Trees Francisco José da Silva Neto , Kátia Priscilla Gomes Morinigo, Nathalia de França Guimarães , Anderson de Souza Gallo , Maicon Douglas Bispo de Souza, Rubismar Stolf, and Anastácia Fontanetti Research Article (8 pages), Article ID 7909467, Volume 2018 (2018)

Propositions on the Optimal Number of Q-Graders and R-Graders Lucas Louzada Pereira , Rogério Carvalho Guarçoni, Gustavo Soares de Souza, Dério Brioschi Junior, Taís Rizzo Moreira , and Carla Schwengber ten Caten Research Article (7 pages), Article ID 3285452, Volume 2018 (2018)

Influence of Solar Radiation and Wet Processing on the Final Quality of Arabica Coffee Lucas Louzada Pereira , Rogério Carvalho Guarçoni, Wilton Soares Cardoso ,RenatoCôrreaTaques, Taís Rizzo Moreira , Samuel Ferreira da Silva, and Carla Schwengber ten Caten Research Article (9 pages), Article ID 6408571, Volume 2018 (2018) Hindawi Journal of Food Quality Volume 2018, Article ID 9805635, 1 page https://doi.org/10.1155/2018/9805635

Editorial Coffee Quality: Cultivars, Blends, Processing, and Storage Impact

Gabriel Henrique Horta de Oliveira ,1 Ana Paula Lelis Rodrigues de Oliveira ,1 Fernando Mendes Botelho,2 Pedro Casanova Treto ,3 and Silvia de Carvalho Campos Botelho 4

1Instituto Federal de Educação, Ciˆencia e Tecnologia do Sudeste de Minas Gerais, Campus Manhuaçu, Manhuaçu, MG, Brazil 2Agrarian and Environmental Sciences Institute, Universidade Federal de Mato Grosso, Sinop, MT, Brazil 3Research Institute of Engineering, Universidad de Costa Rica, San Pedro de Montes de Oca, San Jose´, Costa Rica 4Embrapa Agrossilvipastoril, Sinop, MT, Brazil

Correspondence should be addressed to Gabriel Henrique Horta de Oliveira; [email protected]

Received 9 May 2018; Accepted 9 May 2018; Published 19 June 2018

Copyright © 2018 Gabriel Henrique Horta de Oliveira et al. &is is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Coffee is one of the most important agricultural products in stage, ripe cherries’ Brix degree, percentage of black, unripe, the world, being produced in different countries and under and insect damaged beans, bean size, and beverage quality. several conditions, which leads to an enormous variety of &e best cup quality was obtained in coffee beans coming beverages, according to cultivar selection, types of blends, from coffee trees closer to shaded trees. processing technologies, and storage procedures, among &e fourth paper of this special issue analyzed the op- other features. To spread and foment the discussion re- timal number of Q-graders and R-graders on the sensory garding these sources of variations to coffee drink among analysis consistency for specialty coffees. &e authors producers, industry, and consumers, some important issues indicated that the use of 6 tasters is sufficient to conduct must be addressed such as analysis of residues and micro- sensorial analysis following SCA and BSCA protocol for nutrients in coffee grain, emerging technologies like con- coffees in the Arabica group, as well as 6 tasters for coil and trolled fermentations, intercropping and solar radiation, Conilon coffees. Additional tasters did not improve the methods of sensory analysis, and so on. In this special issue sensorial analysis. &e fifth paper researched the influence of on coffee quality, we have invited a few papers that address solar radiation and wet processing on the final quality of such issues. arabica coffee, being indicated that water fermentation and &e first paper of this special issue aimed at quantifying shaded region are more likely to provide coffee with higher the ash content and determining the concentration of heavy grades. &e final paper investigated, for two consecutive metals in roasted ground coffee. &ese parameters are im- seasons, the effect of two different applications of boron, portant due to their persistency in the environment, be- copper, and zinc over productivity and cup quality. Ap- coming an indicator for coffee quality. &e second paper plication via foliar spray presented better results than trunk presents the study on the yeast fermentation of green coffee injections, leading to higher productivity and cup quality. beans, which consumers indicated that these coffees did not present negative aroma or flavor and presented higher Gabriel Henrique Horta de Oliveira antioxidant activity than coffee without fermentation. &e Ana Paula Lelis Rodrigues de Oliveira third paper is on the influence of different distances of Fernando Mendes Botelho shading coffee trees on plant height, canopy diameter, pla- Pedro Casanova Treto giotropic branches’ length, yield, coffee fruits’ phenological Silvia de Carvalho Campos Botelho Hindawi Journal of Food Quality Volume 2018, Article ID 7960231, 14 pages https://doi.org/10.1155/2018/7960231

Research Article Boron, Copper, and Zinc Affect the Productivity, Cup Quality, and Chemical Compounds in Coffee Beans

Junia Maria Clemente,1 Herminia Emilia Prieto Martinez ,1 Adriene Woods Pedrosa ,2 Yonara Poltronieri Neves,2 Paulo Roberto Cecon,3 and John Lonfover Jifon4

1Department of Plant Science, Universidade Federal de Viçosa, Avenue PH Rolfs, 36.570-000 Viçosa, MG, Brazil 2Empresa de Pesquisa Agropecua´ria de Minas Gerais, EPAMIG Sudeste, Campus UFV–Cx. P. 216, 36.571-000 Viçosa, MG, Brazil 3Department of Statistic, Universidade Federal de Viçosa, Avenue PH Rolfs, 36.570-000 Viçosa, MG, Brazil 4Texas A&M Agrilife Research and Extension Center, Environmental Plant Physiology, 2415 East Business 83, Weslaco, TX 78596, USA

Correspondence should be addressed to Herminia Emilia Prieto Martinez; [email protected]

Received 8 December 2017; Accepted 8 April 2018; Published 14 May 2018

Academic Editor: Ana P. L. R. De Oliveira

Copyright © 2018 Junia Maria Clemente et al. +is is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Micronutrients perform specific and essential functions in plant metabolism, and their deficiency may lead to metabolic disturbances that affect coffee production and quality beverage. In Brazil, the B, Cu, and Zn are the main micronutrients, and these are provided by soil or foliar fertilization, frequently with low recovery efficiency. +is work objected verifying the feasibility of supplying of B, Cu, and Zn via insertion of tablets in the orthotropic branch of Coffea arabica, as well as to evaluate the coffee plant response in terms of productivity and quality of the beverage. Adult plants received B, Cu, and Zn, each micronutrient alone or combined with the other two, by foliar fertilization or by tablets inserted in the trunk base. +e productivity, cupping quality, and some chemical indicators of beans quality were evaluated in two crop seasons. Boron, copper, and zinc supplied by foliar spray or solid injections in the trunk influenced the chemical composition and quality of the coffee beans, characterized by the cupping test and the levels of caffeine, trigonelline, sucrose, glucose, arabinose, mannose, 3-caffeoylquinic acid, 5-caffeoylquinic acid, polyphenol oxidase activity, and total phenolic compounds. Copper and zinc were equivalent in either form of supply regarding the production and quality of coffee.

1. Introduction methods to accurately assess the chemical characteristics and the quality of beverage. Brazil is the largest world producer of coffee. As a world +e production of bioactive compounds related to de- leader in production and exportation, the country needs to sirable flavor and aroma involves extremely complex attend market requirements, innovating and adopting chemical reactions, in which some mineral nutrients could technologies to produce good-quality types of coffee. +e have a key role. However, until now, little is known about the necessity of offering good coffees is rising due to an increase effect of the mineral nutrients B, Cu, and Zn on the pro- in coffee consumers looking for refined tastes and aromas, duction of chemical compounds that define the good quality which are related to the chemical composition of the coffee of coffee. Most of the studies about mineral nutrition and beans. Flavor and aroma are the main criteria to evaluate coffee fertilization are focused on sources and doses of these beverage quality and also constitute the most important nutrients, in order to optimize the productivity. attributes for consuming coffee [1]. +e cup test is the Boron deficiency is common in most of the Brazilian standard approach to evaluate the flavor and aroma of coffee; soils [3]. In coffee plants, its deficiency has been attributed to however, quite often, it is criticized because of its subjective the natural loss of soil fertility, as well as to the wide use of nature [2]. +erefore, it is essential to explore alternative highly demanding varieties. Boron deficiency in coffee plants 2 Journal of Food Quality can reduce the root system growth and cause the death of is directly involved with the quality of the coffee beverage thin root tips and consequently the decrease in water and [2, 16, 17]. It is established that coffee beans that are mineral absorption. In consequence of that, the plants be- strongly damaged or Cu deficient may have low PPO ac- come sensible to drought and less responsive to fertiliza- tivity and low quality. Carvalho et al. [2] performed pio- tions. As B is highly immobile in the phloem sap [4], its neering works, with physical and chemical evaluations of supply via soil is desirable, although been supplied many processed coffee beans previously classified as “strictly times by foliar sprays, mixed with Cu and Zn. soft,” “soft,” “softish,” “hard,” “rioysh,” and “rio,” and +e Cu deficiency compromises the activity of several verified that the coloration index and PPO allow the enzymes that catalyze oxidative reactions of several metabolic separation of beans with different coffee quality types. routes, especially the plastocyanin, superoxide dismutase, and Chlorogenic acids are the main nonvolatile phenolic polyphenoloxidase [5]. In the soil, Cu is strongly complexed by compounds found in coffee beans and account for 6 to 12% organic matter [6], so the common technique used for sup- of their dried mass [18]. +ey are formed by means of es- plying Cu to coffee plants is through foliar sprays. +e Cu terification of trans-cinnamic acids, such as caffeic acid, deficiency can cause irreversible metabolic disturbances in ferulic acid, and p-coumaric acid with quinic acids [19]. coffee plants and possibly compromise the production of During the roast process, they are strongly degraded gen- chemicals related to the beverage quality. erating acids, lactones, and volatile compounds such as the According to Fageria et al. [7], the lack of Zn impairs the phenil, guaiacol, and 4-vinyl guaiacol [20, 21], which con- world agriculture production, as well as the nutritional quality tribute to the flavor and aroma of coffee, especially for the of grains. +e main symptoms of Zn deficiency are related, in astringency of the beverage; the proanthocyanidins along a still unclearly way, to disturbance in auxin metabolism [5, 8], with the polyphenols also provide an astringent flavor [22]. which plays an essential role in the synthesis of tryptophan, an Within acceptable limits, chlorogenic acids have a positive amino acid precursor of IAA [9]. +e photosynthetic activity effect on the beverage body. of coffee plants is hugely diminished under lack of Zn, given its Caffeine, commonly known as 1,3,7-trimethylxanthine, importance on enzymes involved in carbon fixation [10]. belongs to the methylxanthine class and gives bitterness to Besides this, Zn regulates or makes part of the structure of the coffee taste [19]. several other enzymes involved in protein synthesis and in Trigonelline corresponds to around 1% of raw beans, and nitrogen metabolism [5, 11]. In clayey acid soils, several re- it is one of the precursors of aroma in coffee and undergoes actions are responsible for the low Zn availability, which degradation of up to 90% during roasting, forming mainly together with the low mobility of Zn in the coffee plant phloem niacin, pyridines, and some pyrroles; the lower the trig- drives to foliar fertilizations in such conditions [12, 13]. Also, onelline content in the beans, the lower the quality of the like other Brazilian coffee producer regions, Zona da Mata is coffee beverage [23]. It is worth to highlight that the more a mountain region in which manually done foliar sprays are drastic one is the roast process in which the lower levels of time-consuming and costly. trigonelline will be found in the samples [24]. Coffees of superior quality are those that have chemical Production of secondary metabolism metabolites, such compounds responsible for the flavor and aroma such as as polyphenols, caffeine, trigonelline, alcohols, and alde- caffeine, trigonelline, aldehydes, furans, ketones, sugars, hydes, depends on the primary metabolism and its catabolic proteins, amino acids, pyrroles, pyridines, pyrazines, oxa- reactions that produce energy and the carbonic skeleton, zoles, carboxylic acids, fatty acids, and phenolic compounds such as sucrose. +erefore, if any factor affects photosyn- in an equilibrated proportion to obtain good body, acidity, thates production during the fruits development, then it can and smoothness of the beverage. also affect negatively the quality of beverage [25]. When in the presence of microorganisms or under +is work is aimed at evaluating the production of anaerobic conditions, the sugars present in the coffee mu- bioactive compounds and the quality of raw coffee beans cilage can be fermented and produce alcohols that may be harvested from plants fertilized with B, Cu, and Zn via foliar broken down successively in acetic, lactic, propionic, and sprays or solid injections of salts in the trunk and correlates butyric acids. Pinto et al. [14] studying the quality of beans these variables with the nutritional status of the plants. used to prepare espresso coffee observed that the low-quality ones, such as the types “rio” and “rioysh,” presented higher 2. Materials and Methods acidity than the types “strictly soft” and “soft.” Electrical conductivity (EC) or the leached potassium +ree experiments were performed in July, in a field crop (LK) has been used in researches as consistent indicators of area of the Universidade Federal de Viçosa, Viçosa, Minas cellular membrane integrity. +ey are accessory attributes Gerais State, Brazil, using an adult orchard of Coffea used preferentially to differentiate beverages of the same arabica L. cv. Catua´ı IAC-99 and have been conducted and class, having little adequacy as a unique mean of differen- evaluated during two crop seasons. +e experimental field tiation. Grains of bad quality commonly present higher EC is located at 20°45 south and 42°51 west, at 541 m above the and less organization and cellular structuring than the good sea level. +e soil of the experimental field is classified as ones, showing that EC is a strong indicator of membrane and a red-yellow latosol and the climate is classified as Cwa, cell wall damage [15]. according to the Ko¨ppen classification, with annual Polyphenoloxidase (PPO) is a cupric enzyme linked to temperature and precipitation averages of 19.4°C and the cellular membranes, and as discussed in the literature, it 1221.4 mm, respectively. Journal of Food Quality 3

2.1. Treatments and Experimental Design. In the first ex- Table 1: Numerical scores for the coffee cupping test. periment, the following treatments were performed to evaluate Taste Nota the boron effect: control without B supply, foliar sprays with Strictly soft (specialty coffee) ≥87 boric acid at 0.4%, injection of tablets containing B salts at the Soft 80–86 base of the trunk, injection of tablets containing B + Cu salts at Softish 74–79 the base of the trunk, injection of tablets containing B + Zn Hard ≤74 salts at the base of the trunk, and injection of tablets containing B + Cu + Zn salts at the base of the trunk. +e second experiment, to evaluate the copper effect, was +e content of Cu and Zn was determined by atomic ab- performed in the same way as the first, receiving the following sorption spectrophotometry [28] in the extract of the nitric- treatments: control without Cu supply, foliar sprays with perchloric acid digestion [29]. copper sulphate at 0.4%, injection of tablets containing Cu salts at the base of the trunk, injection of tablets containing Cu + B salts at the base of the trunk, injection of tablets containing 2.2. Evaluations Cu + Zn salts at the base of the trunk, and injection of tablets 2.2.1. Production. +e harvesting of the four usable plants of containing Cu + B + Zn salts at the base of the trunk. the plot was carried out, when the plants had approximately +e third experiment, to evaluate the zinc effect, was 95% coffee cherries. +e coffee cherries were handpicked and performed like the other two, receiving the following treat- dried on a bench in a greenhouse until achieving 11% ments: control without Zn supply, foliar sprays with zinc moisture content. After drying, they were hulled and used sulphate at 0.4%, injection of tablets containing Zn salts at the for the chemical analysis. base of the trunk, injection of tablets containing Zn + B salts at the base of the trunk, injection of tablets containing Zn + Cu salts at the base of the trunk, and injection of tablets containing 2.2.2. Cupping Quality. +e cupping test was performed by Zn + B + Cu salts at the base of the trunk. professional tasters, using the CoE (cup of excellence) All the experiments were assigned as randomized blocks method (Table 1). Each attribute (clean beverage, sweetness, with 5 replications. Each plot was composed of 18 plants acidity, body, taste, flavor, and reminiscent taste) received distributed in three rows 3 m apart and with 1 m between a score based on the taste intensity exhibited by the samples, plants in a row. +e four central plants constituted the useful according to the Brazilian official method plus the grades plot. We worked with 30 total plots in each experiment. described by the SCAA method for specialty coffees [30]. +e tablets of B and Zn were prepared at the Laboratory of Civil Engineering, Universidade Federal de Viçosa, using a hydraulic press with a force of 0.5 tons. +e copper was 2.2.3. Chemical Analysis of the Coffee Beans. Total sugars, supplied through capsules without any compression of the nonreducing sugars, coloration index, total titratable acidity, salts, due to the difficulty of these in forming compact mass pH, electrical conductivity, and leached potassium were with excipient agents. +e tablets were implanted into the evaluated in beans harvested in the crop season 2010/2011. orthotropic branch of the coffee tree at 10 cm above the Coloration index, leached potassium, total titratable acidity, ground. pH, electrical conductivity, caffeine, trigonelline, total Considering that usually coffee foliar sprays are done with phenolic compounds, sucrose, glucose, mannose, arabinose, a volume of 400 L·ha−1 and that the orchard had 3333 plants galactose, proanthocyanidin, 3-caffeoylquinic acid, 4-caf- per hectare, we used 120 mL of spraying solution per plant, feoylquinic acid, 5-caffeoylquinic acid, and PPO activity totaling 480 mL per plot. All nutrients were sprayed by means were evaluated in beans harvested in the crop season of handheld sprayers with cone-filled nozzles, and we added 2011/2012 as described below. All the extractions and 1 mL·L−1 of the adhesive adjuvant to the solution. Plastic readings were performed in duplicates. curtains were placed between the rows to prevent drift. Total sugars and reducing sugars were extracted by the In each crop season, three foliar sprays were applied Lane–Enyon method as described by the Association of between September and February. +e liming and fertil- Official Analytical Chemists [31] and determined by the ization with nitrogen, phosphorus, and potassium were Somogyi technique, adjusted by Nelson [32]. +e non- performed based on soil analysis, and the expected pro- reducing sugars were determined by the difference between ductivity, following the recommendations of [26]. total and reducing sugars. +e coloration index was de- In order to determine the nutritional status of the plants termined by Singleton [33], adapted for coffee. +e results of subjected to the different treatments, in the two crop sea- total sugar were expressed in percentage (%) and the col- sons, coffee leaves were taken from the third or fourth nodes oration index in DO 425 nm, respectively. and counted from the apex to the base of plagiotropic Total titratable acidity and the pH were determined as branches, at a median height in the canopy and in the period described by the Association of Official Analytical Chemists between flowering and the first rapid expansion of the fruits. [31]. Already, electrical conductivity was determined according +e leaves were washed in deionized water and dried in an to the method described by Loeffler et al. [34], and the leached oven with forced air at 70°C, until constant weight. potassium was determined using a flame photometer, as de- Boron content was determined using the azomethine-H scribed by Prete [35]. +e results of total titratable acidity, method after dry digestion of the plant material [27]. electrical conductivity, and leached potassium were expressed 4 Journal of Food Quality

Table 2: Contents of B, Cu, and Zn (mg·kg−1) of index leaves of 2.2.4. Statistics. Data were submitted to variance analysis, coffee plants that received B, Cu, and Zn as solid injections or foliar and the means were compared by Dunnett’s test at 10% of sprays (FSs). probability in the program SAEG 9.1 [42]. +e treatment Crop season 2010/2011 2011/2012 without B, Cu, and Zn was considered as the first control Boron because it is the control of all treatments, and the sprayed WB 24.08∗ 27.46 treatment was considered as the second control because it is FS (control) 35.04+ 30.35 the usual form to supply B, Cu, and Zn to coffee plants. B 35.51+ 50.41∗,+ B + Cu 35.04+ 64.65∗,+ + ,+ 3. Results and Discussion B + Zn 33.26 72.31∗ ∗ ∗,+ B + Cu + Zn 23.94 78.64 3.1. Boron Content in Index Leaves. In the crop season CV (%) 6.59 13.19 2010/2011, the index leaves of the coffee plants that received Copper WCu 9.64∗ 5.24∗ B as foliar spray or solid injections with B, B + Cu, or B + Zn FS (control) 13.51+ 13.14+ presented higher contents of B than the control treatment Cu 17.06∗,+ 18.62∗,+ without boron (WB). For this same crop season, the ex- B + Cu 14.94+ 15.79∗,+ periments evaluating tablets of Cu and Zn behaved the same Cu + Zn 18.96∗,+ 19.41∗,+ way compared to the control treatments without Cu and B + Cu + Zn 16.46∗,+ 16.52∗,+ without Zn (WCu-WZn; Table 2). CV (%) 8.25 6.15 +e content of B in the leaves of the treatments with B, Zinc B + Zn, and B + Cu was statistically similar to that in the WZn 6.5∗ 4.68∗ + + sprayed control treatment, and the treatment with B + Cu + Zn FS (control) 10.06 7.42 was statistically lower in the first crop season (Table 2). Zn 10.85+ 11.58∗,+ +e content of Cu of the treatments with Cu, Cu + Zn, B + Zn 10.63+ 10.89∗,+ Cu + Zn 11.75∗,+ 13.66∗,+ and B + Cu + Zn was statistically higher than that observed B + Cu + Zn 9.85+ 9.98∗,+ in the sprayed control, suggesting fast release of the nutrient CV (%) 6.25 16.31 in the treatments that received Cu as solid injections. For the WB, WCu, and WZn: control treatments, without application of B, Cu, and Zn content in leaves, only the Cu + Zn treatment was sig- Zn, respectively; FS: foliar spray with boric acid, copper sulphate, and zinc nificantly higher than the sprayed control (Table 2). sulphate (0.4%); B: trunk injection of tablets containing B salts; Cu: trunk Considering the sufficiency ranges of 29 to 52 mg·kg−1 injection of tablets containing Cu salts; Zn: trunk injection of tablets for B, 13 to 29 mg·kg−1 for Cu, and 6 to 12 mg·kg−1 for Zn as containing Zn salts; B + Cu: trunk injection of tablets containing B and Cu determined by Martinez et al. [43] for the region of Viçosa, salts; B + Zn: trunk injection of tablets containing B and Zn salts; Cu + Zn: trunk injection of tablets containing Cu and Zn salts; B + Cu + Zn: trunk only the plants of the control treatment (WB-WCu) were injection of tablets containing B, Cu, and Zn salts; ∗mean values are sta- deficient in B and Cu and the plants of the treatment with tistically different from those of the control treatment (FS) at the 10% B + Cu + Zn were deficient in B. + significance level, according to Dunnett’s test; mean values are statistically In the crop season 2011/2012, solid injections of B, Cu, different from those of the control treatment (WB-WCu-WZn) at the 10% and Zn in the trunk, pure or combined with two or three significance level, according to Dunnett’s test. elements, resulted in higher contents of these nutrients in leaves than the treatments WB, WCu, and WZn. In the case of B, the concentrations attained could be considered toxic −1 −1 −1 −1 in mL of NaOH 100 g of the sample, µS·cm ·g , and g·kg , to the plants, while in comparison to the sprayed treatment, respectively. it can be noted that the content of all treatments that re- +e polyphenoloxidase (PPO) activity was determined as ceived solid injections of B, Cu, and Zn was significantly described by Ponting and Joslyn [36], using the sample extract higher (Table 2). without DOPA as the blank, and the results were expressed in In all treatments, except the treatments WB, WCu, and U/min/g of the sample. Chlorogenic acids were extracted WZn, the concentrations of Cu and Zn were considered according to Farah et al. [37] and Trugo and Macrae [18], and adequate according to the method described by Martinez the results were expressed in percentage (%). et al. [43]. +e results suggest that both forms, in both crop Caffeine was determined according to the method de- seasons, were efficient to increase the contents of B, Cu, and scribed by Mazzafera et al. [38] with additional modifica- Zn in index leaves of coffee plants, even with the necessity to tions as described by Vitorino et al. [39], the trigonelline was review the composition and doses of B salts. determined according to the method described by Vitorino et al. [39], and phenolic compounds were determined by the Folin–Denis method as described by the Association of 3.2. Production. In the first crop season, there were no Official Analytical Chemists [31]. +e results were expressed statistically significant differences in coffee production. in percentage (%). Plants in the B experiment yielded 3.59 kg of cherries per +e proanthocyanidins were determined as described by plant (2992 kg·ha−1 of processed coffee), while plants in the Hagerman et al. [40], and sucrose, glucose, mannose, galactose, Cu and Zn experiments produced on average 3.61 and and arabinose were determined as described by Sluiter et al. 3.54 kg of coffee cherries per plant (3.005 kg·ha−1 and [41]. +e results were expressed in percentage (%). 2947 kg·ha−1 of processed coffee, resp.). +is result could be Journal of Food Quality 5

Table 3: Coffee cherry production of coffee plants submitted to the salts supplied Cu in adequate amounts, but the most in- fertilization via solid salts injections in the trunk and foliar sprays teresting finding was the good combination with B and Cu in with B, Cu, and Zn. the same tablet, since the treatments containing only Cu, Treatments Production Cu + Zn, and B + Cu + Zn and the sprayed treatment were 2010/2011 2011/2012 statistically equal compared to the treatment WCu. Later, Boron maybe the plants were slightly sensitive to high concen- WB 3.47 4.55 trations of Cu, as can be concluded taking into account the FS 3.57 5.47 Cu content in the index leaves. B 3.83 4.30 For the Zn experiment, there was no significant effect of B + Cu 3.74 6.15∗,+ Zn on production, in the two crop seasons, even with the ,+ B + Zn 3.47 5.59∗ variations in the content of Zn in index leaves, with the mean B + Cu + Zn 3.45 3.93 values of the production in the second crop year being 4.99 Means 3.59 5.00 (4157 kg·ha−1 of processed coffee) (Table 3). CV (%) 22.88 26.66 +e high level of probability used can be considered Copper WCu 3.47 4.55 acceptable for coffee experiments conducted in commercial FS 3.57 5.47 orchards because the experimental conditions are very Cu 3.99 5.56 heterogeneous and each coffee plant of the plant population B + Cu 3.74 6.15∗,+ has great variability. Cu + Zn 3.42 5.15 According to Brown and Shelp [44], the B moves in the B + Cu + Zn 3.45 3.93 form of complexes with polyols (sugar-alcohol), and coffee Means 3.61 5.13 has a large amount of mannitol, but there is little in- CV (%) 22.85 20.38 formation on the distribution of these polyols. Brown and Zinc Hu [45] observed that, in coffee plants, the B is immobile in WZn 3.47 4.55 phloem because of the low capacity of forming stable FS 3.57 5.47 complexes with sucrose; therefore, the foliar sprays correct Zn 3.83 5.24 B + Zn 3.47 5.59 the deficiency only in the leaves that received the fertilizers, Cu + Zn 3.42 5.15 and the leaves that grow after fertilizers application will B + Cu + Zn 3.45 3.93 present low B concentrations, demanding a greater number Means 3.54 4.99 of applications. CV (%) 28.56 27.03 Santinato et al. [46] working with high doses of boric WB, WCu, and WZn: control treatments, without application of B, Cu, and acid applied in the soil observed that, despite the plant did Zn, respectively; FS: foliar spray with boric acid, copper sulphate, and zinc not present toxicity symptoms, excessive doses caused re- sulphate (0.4%); B: trunk injection of tablets containing B salts; Cu: trunk duction of 600 kg·ha−1 of processed coffee. injection of tablets containing Cu salts; Zn: trunk injection of tablets Positive correlations between B availability in soil and containing Zn salts; B + Cu: trunk injection of tablets containing B and Cu salts; B + Zn: trunk injection of tablets containing B and Zn salts; Cu + Zn: the harvest index were found by Lima Filho and Malavolta trunk injection of tablets containing Cu and Zn salts, B + Cu + Zn: trunk [47] for Coffea arabica cv. Catua´ı Amarelo. +ey have ob- injection of tablets containing B, Cu, and Zn salts; ∗mean values are sta- served a great correlation among the harvest index and B tistically different from those of the control treatment (FS) at the 10% content in leaves, branches length, and number of leaves and significance level, according to Dunnett’s test; +mean values are statistically different from those of the control treatment (WB-WCu-WZn) at the 10% branches and low correlation among the harvest index and significance level, according to Dunnett’s test. dry matter of the roots, stem, branches, and leaves. +ese variables are important for coffee production because, according to Rena and Maestri [48], the vertical growth of coffee plant determines the formation of nodes, and from due to the fact that fruits are produced in nodes formed in buds of these nodes emerge plagiotropic branches, in which the previous growing season; therefore, nodes in which the nodes will develop leaves and inflorescences. +erefore, the fruiting occurred were already formed prior to treatment flowering depends on the branches growth, the number of applications in this study (Table 3). nodes, and numbers of leaves, since it is to verify that many In the crop season 2011/2012, there was a significant nodes without leaves do not flower. difference in coffee production among the treatments at 10.9 % According to Santinato et al. [49], 6 to 12 applications of probability (Table 3), with the productions of the treatments per year with solutions at the concentration of 0.25% organic containing B (B + Cu and B + Zn) 35.01% and 22.82% greater boron (10% B) not coinciding with the flowering provided than that in the treatment WB. Such differences correspond to high productivity and maintained good correlation between 1328 kg·ha−1 and 866 kg·ha−1 of processed coffee, respectively, B content in leaves and production, although there was no even with the index leaves presenting excessive concentrations significant difference between the treatments regarding the of the nutrient (Tables 2 and 3). B content in leaves. Barros et al. [50] observed that boric acid For the Cu experiment, the production of the treatment application at 0.3% twice a year resulted in productions with B + Cu was statistically different, considering the only 8% higher than the control treatment without B ap- treatment without Cu or the sprayed treatment as controls plication. On the other hand, Lima Filho and Malavolta (Table 3). +ese results suggest that tablets containing B + Cu [47] and Barros et al. [50] reported that the increase in B 6 Journal of Food Quality concentration not always provides an increase in coffee Table 4: Cupping test of coffee beans harvested from plants productivity. submitted to fertilization via solid salts injections and foliar spray According to Andrade [51], in adult coffee plants re- with B, Cu, and Zn. ceiving sprays containing Cu, high Cu content present in leaf Treatments 2010/2011 2011/2012 did not reduce production possibly because Cu is located on Cupping test the leaf surface or even because the element remains largely Boron in the cuticle not reaching the cytoplasm. Loneragan [52] WB 82.4 82.7 states that Cu movement into the plants depends on its FS 84.8 83.9 concentration. During the initial stages of growth, Cu in B 84.4 72.8 excess causes reduction on the branching, thickening, and B + Cu 83.4 83.25 abnormal coloration of rootlets [53]. B + Zn 83.6 79.35 Regarding Zn in a field experiment, Guimarães et al. [54] B + Cu + Zn 83.8 80.4 observed an increase of 60 to 360 kg·ha−1 of processed coffee CV (%) 3.76 8.20 with Zn supplementation by foliar spray, followed by the Copper · −1 WCu 82.4 82.7 increase in the concentration of Zn from 8 to 21 mg kg in FS 84.8 83.9 index leaves. In turn, Lima Filho and Malavolta [55] proved Cu 85.0 81.5 the positive interaction between B and Zn studying the dry B + Cu 83.4 83.25 matter production in seedlings of coffee varieties, and when Cu + Zn 82.0 74.6∗,+ the nutrients were supplied together, the dry matter pro- B + Cu + Zn 83.8 80.4 duction was 21% higher. CV (%) 4.27 5.81 In a field experiment performed in the same conditions Zinc of that of the experiment reported here, Martinez et al. [56] WZn 82.4 82.7 studied the effect of Zn on the production and on some FS 84.8 83.9 quality attributes of coffee beans and did not observe effect of Zn 84.2 78.9 B + Zn 83.6 79.35 the nutrient on the production; however, there was an effect Cu + Zn 82.0 74.6 of Zn on beans size. Moreover, plants supplemented with Zn B + Cu + Zn 83.8 80.4 had the highest percentage of exportable grains, retained in CV (%) 4.34 10.17 the sieves 17 and 18. Still according to these authors, there WB, WCu, and WZn: control treatments, without application of B, Cu, was no significant effect of Zn on the cup quality of coffee and Zn, respectively; FS: foliar spray with boric acid, copper sulphate, beans; however, the score related to the cup quality of the and zinc sulphate (0.4%); B: trunk injection of tablets containing B salts; beans harvested from plants that did not receive Zn was 60, Cu: trunk injection of tablets containing Cu salts; Zn: trunk injection of while in the beans harvested from plants that received Zn, it tablets containing Zn salts; B + Cu: trunk injection of tablets containing B and Cu salts; B + Zn: trunk injection of tablets containing B and Zn salts; was 72.5. Cu + Zn: trunk injection of tablets containing Cu and Zn salts; B + Cu + Zn: In the same orchard, Neves et al. [57] studied the effect of trunk injection of tablets containing B, Cu, and Zn salts; ∗mean values are Zn, supplied by trunk injection and foliar sprays, on the statistically different from those of the control treatment (FS) at the 10% production and on some attributes of quality. +e cumu- significance level, according to Dunnett’s test; +mean values are statistically lative production of two crop seasons for treatments that different from those of the control treatment (WB-WCu-WZn) at the 10% significance level, according to Dunnett’s test. received tablets of Zn inserted in the trunk and the treatment without Zn was 11,292 and 7810 kg·ha−1 of processed coffee, respectively. +e difference among them was 3482 kg·ha−1, which corresponds at 30.9%. Still according to the authors, that when in appropriate concentrations certainly would the beans were classified as “hard” type, and there was no influence positively the route of production of com- significant effect of Zn on coffee bean quality evaluated by pounds associated with desirable flavors and aromas the cup test. (Tables 2 and 4). +ere was no effect of Zn on the cupping test, reaching scores of 83.46 and 79.97 in the two crop seasons, re- 3.3. Cupping Quality. +ere was no significant effect of B on spectively, and the beans would be classified as “soft” the cupping test in both years, with the overall mean grades (Table 4). In spite of the fact that the precision and validity of 83.73 and 80.4 in the respective assessed years; in general, of the cupping are much discussed because of its subjective the coffee beans were classified as “soft” (Table 4). nature and the limitation of tasters abilities, this result Comparing tablets with the treatment WCu and the points out to the major importance of the postharvest sprayed treatment, only the treatment containing Cu procedures than the mineral nutrition of the plant on + Zn had statistically low scores in the crop season coffee quality. 2011/2012, evidencing the effect of the way of Cu supply and the effect of the nutrient on the cupping quality. As previously reported, the index leaves of the plants sub- 3.4. Electrical Conductivity, Leached Potassium, and Color- jected to this treatment had slightly high Cu content. In ation Index in the Crop Season 2010/2011. In the first crop case of Cu and Zn tending to the excess quantity, there is season, there was no effect of B, Cu, and Zn supplied as a negative interaction with other cationic micronutrients tablets on CI, TTA, pH, EC, KL, TS, and RS of the beans Journal of Food Quality 7

Table 5: Coloration index (CI, DO 425 nm), total titratable acidity Table 6: Coloration index (CI, DO 425 nm), total titratable acidity (TTA, mL NaOH 100 g−1), pH, electrical conductivity (EC, (TTA, mL of NaOH 100 g−1), pH, electrical conductivity (EC, µS·cm−1·g−1), leached potassium (KL, g·kg−1), total sugars (TS, %), µS·cm−1·g−1), and leached potassium (KL, g·kg−1) of coffee beans of and reducing sugars (RS, %) in coffee beans of Coffea arabica Coffea arabica treated with different forms of B, Cu, and Zn supply, treated with different forms of B, Cu, and Zn supply, in the crop in the crop season 2011/2012. season 2010/2011. Crop season 2011/2012 Crop season 2010/2011 Boron Boron Treatments CI TTA pH EC KL Treatments CI TTA pH EC KL TS RS WB 0.51 49.41 5.72 57.52 2.15 WB 0.81 11.2 5.61 41.39 1.29 11.73 0.216 FS (control) 0.76 48.42 5.63+ 50.18 2.01 FS 0.89 10.4 5.60 42.71 1.40 11.41 0.197 B 0.74 41.50 5.63+ 56.77 2.27 B 0.84 10.0 5.60 41.36 1.30 11.63 0.212 B + Cu 0.37∗ 40.52 5.68 60.98 2.05 B + Cu 0.93 10.4 5.58 41.84 1.29 12.09 0.199 B + Zn 0.66 45.46 5.69 58.83 1.86 B + Zn 0.91 10.8 5.50 41.09 1.23 9.70 0.279 B + Cu + Zn 0.41∗ 50.40 5.74∗ 59.27 2.02 B + Cu + Zn 0.93 11.2 5.61 38.87 1.23 11.28 0.264 CV (%) 31.89 31.12 0.86 13.47 14.28 CV (%) 12.03 20.05 1.35 14.26 12.67 13.91 34.39 Copper Copper WCu 0.51 49.41 5.72 57.52 2.15 Treatments CI TTA pH EC KL TS RS FS (control) 0.76+ 48.42 5.63 50.18 2.01 WCu 0.81 11.2 5.61 41.39 1.29 11.73 0.216 Cu 0.50∗ 59.29 5.64 56.92 1.93 FS 0.89 10.4 5.61 42.71 1.40 11.41 0.197 Cu + B 0.37∗ 40.52 5.68 60.98 2.05 Cu 0.87 10.4 5.60 41.98 1.36 11.48 0.221 Cu + Zn 0.50∗ 48.42 5.70 62.73 2.12 Cu + B 0.93 10.4 5.58 41.84 1.29 12.09 0.199 B + Cu + Zn 0.41∗ 50.40 5.74 59.27 2.02 Cu + Zn 0.88 11.6 5.58 38.14 1.18 12.20 0.211 CV (%) 29.25 31.61 1.26 12.79 14.15 B + Cu + Zn 0.93 11.2 5.61 38.87 1.23 11.28 0.264 Zinc CV (%) 10.76 24.48 0.75 13.8 13.67 14.16 20.22 WZn 0.51 49.41 5.71 57.52 2.15 Zinc FS (control) 0.76 48.42 5.63 50.18 2.01 Treatments CI TTA pH EC KL TS RS Zn 0.52 42.49 5.65 58.29 2.36 WZn 0.80 11.2 5.61 41.39 1.29 11.73 0.216 Zn + B 0.66 45.46 5.69 58.83 1.86 FS 0.89 10.4 5.60 42.71 1.40 11.41 0.197 Zn + Cu 0.50 48.42 5.70 62.73 2.12 Zn 0.90 11.2 5.62 41.09 1.32 12.13 0.209 B + Cu + Zn 0.41 50.40 5.74 59.27 2.02 Zn + B 0.91 10.8 5.50 41.09 1.23 9.70 0.279 CV (%) 31.29 30.14 1.17 15.86 15.91 Zn + Cu 0.88 11.6 5.58 38.14 1.18 12.20 0.211 WB, WCu, and WZn: control treatments, without application of B, Cu, and B + Cu + Zn 0.93 11.2 5.61 38.87 1.23 11.28 0.264 Zn, respectively; FS; foliar spray with boric acid, copper sulphate, and zinc CV (%) 11.06 17.27 1.45 16.7 14.13 17.24 35.64 sulphate (0.4%); B: trunk injection of tablets containing B salts; Cu: trunk WB, WCu, and WZn: control treatments, without application of B, Cu, injection of tablets containing Cu salts; Zn: trunk injection of tablets and Zn, respectively; FS: foliar spray with boric acid, copper sulphate, containing Zn salts; B + Cu: trunk injection of tablets containing B and Cu and zinc sulphate (0.4%); B: trunk injection of tablets containing B salts; salts; B + Zn: trunk injection of tablets containing B and Zn salts; Cu + Zn: Cu: trunk injection of tablets containing Cu salts; Zn: trunk injection of trunk injection of tablets containing Cu and Zn salts; B + Cu + Zn: trunk ∗ tablets containing Zn salts; B + Cu: trunk injection of tablets containing B injection of tablets containing B, Cu, and Zn salts; mean values are statistically different from those of the control treatment (FS) at the 10% and Cu salts; B + Zn: trunk injection of tablets containing B and Zn salts; + Cu + Zn: trunk injection of tablets containing Cu and Zn salts; B + Cu + Zn: significance level, according to Dunnett’s test; mean values are statistically trunk injection of tablets containing B, Cu, and Zn salts; ∗mean values are different from those of the control treatment (WB-WCu-WZn) at the 10% statistically different from those of the control treatment (FS) at the 10% significance level, according to Dunnett’s test. significance level, according to Dunnett’s test; +mean values are statistically different from those of the control treatment (WB-WCu-WZn) at the 10% significance level, according to Dunnett’s test. Zinc did not influence the CI, TTA, pH, EC, and KL in this crop season (Table 6). According to Carvalho et al. [2], the coloration index allows separation of different types of coffees, such as compared to the treatments WB, WCu, and WZn or the “rioysh” and “rio,” that are not acceptable drink with col- sprayed treatments as controls (Table 5). oration indexes lower than 0.650 DO 425 nm. For those In the second crop season, there was no effect of B on the classified as “hard” (acceptable), “soft,” “softish” (fine), and TTA, EC, and KL. Only the treatments with B + Cu + Zn and “strictly soft” (extra fine), the coloration indexes would be B + Cu differed from the sprayed treatment but did not differ equal or greater than 0.650 DO 425 nm. from the treatment without B, indicating the effect of the Results obtained during the second crop season of way of B supply on this variable but not the effect of the evaluation suggest that the coloration index may not be nutrient (Table 6). a good indicator of quality as assessed by the cupping test. +ere was no effect of Cu on TTA, pH, EC, and KL. +e Although color indices were very similar, cup quality scores treatments that received Cu, Cu + B, Cu + Zn, and B + Cu for beans from trees with spray treatments were 15.24% + Zn via solid injections differed from the sprayed treatment, higher than those with trunk insertion of B tablets (Tables 4 and considering the treatment WCu as control, only the and 6). sprayed treatment was different, indicating the effect of Cu +e coloration index of the sprayed treatment is in and the way of Cu supply on CI (Table 6). agreement with that reported by [2] in which the coffees of 8 Journal of Food Quality best quality were darker and the dark coloration attributed Table 7: Caffeine (Caf, %), trigonelline (Trig, %), sucrose (Suc, %), to the formation of essential compounds to develop desirable glucose (Glu, %), galactose (Gal, %), arabinose (Ara, %), and flavors and aromas (Table 6). mannose (Man, %) of coffee beans harvested from plants submitted On the other hand, the treatments containing B + Cu and to the fertilization via solid salts injections in the trunk or foliar B + Cu + Zn presented the coloration index quite low (0.368 sprays with B, Cu, and Zn, in the crop season 2011/2012. and 0.411 DO 425 nm, resp.) and high cupping quality scores Treatments Caf Trig Suc Glu Gal Ara Man (83.25 and 80.4, resp.), being, in this case, classified as Boron “softish,” probably because of the positive effects of Cu and WB 1.01∗ 0.83∗ 5.14∗ 0.18∗ 0.10 0.03 0.14 Zn, contradicting the findings of Carvalho et al. [2]. Correaˆ FS 1.51+ 0.96+ 6.45+ 0.26+ 0.14 0.03 0.15 et al. [58] also reported that higher CI could be attributed to B 1.51+ 0.98+ 6.61+ 0.28+ 0.11 0.03 0.15 + + + + ,+ the occurrence of biochemical alterations and oxidative B + Cu 1.46 0.96 6.45 0.28 0.11∗ 0.05∗+ 0.17∗ + + + ∗ ∗ ∗,+ reactions caused by the dry conditions or inadequate B + Zn 1.50 0.99 6.81 0.35 + 0.16 0.04 0.15 B + Cu storage. 0.93∗ 0.85∗ 5.10∗ 0.24+ 0.11 0.03 0.14 Martinez et al. [56] studying Zn effect on the production + Zn CV (%) 5.85 5.54 6.04 14.32 42.61 19.23 6.59 and cupping quality of coffee did not observe Zn effect on the Copper coloration index, with the mean values being 0.95 DO WCu 1.01∗ 0.83∗ 5.14∗ 0.18∗ 0.10∗ 0.03 0.14∗ 425 nm, which is in agreement with the results found by FS 1.51+ 0.96+ 6.45+ 0.26+ 0.14+ 0.03 0.15+ Lima et al. [59] for good coffees and, also, with the means Cu 1.50+ 0.99+ 6.01+ 0.29+ 0.10∗ 0.02 0.16∗,+ found in this work. B + Cu 1.46+ 0.96+ 6.45+ 0.28+ 0.11∗ 0.05∗+ 0.17∗,+ Cellular membrane damage and the subsequent loss of Cu + Zn 1.46+ 0.97+ 6.53+ 0.30+ 0.12 0.03 0.16+ B + Cu permeability control were proposed by Heydecker Vigour 0.93∗ 0.85∗ 5.10∗ 0.24 0.11∗ 0.03 0.14∗ [60] and Harrington [61] as the early step in the seed de- + Zn terioration process. According to Amorim [16], since the CV (%) 6.16 4.63 6.14 16.78 17.06 19.39 4.05 leached potassium is proportional to the loss of bean quality, Zinc WZn 1.01∗ 0.83∗ 5.14∗ 0.18∗ 0.10 0.03 0.14∗ it can be observed that the loss of membrane permeability FS 1.51+ 0.96+ 6.45+ 0.26+ 0.14 0.03 0.15+ caused damage in coffee beans. Zn 1.55+ 0.99+ 6.62+ 0.31+ 0.10 0.04∗,+ 0.16+ Malta et al. [62] studying some attributes related to the B + Zn 1.50+ 0.99+ 6.81+ 0.35∗,+ 0.16 0.04∗,+ 0.15 quality of different coffee varieties noted that Catua´ı Ver- Cu + Zn 1.46+ 0.97+ 6.53+ 0.30+ 0.12 0.03 0.16+ −1 −1 melho presented an electrical conductivity of 104 µS·cm ·g , B + Cu 0.93∗ 0.85∗ 5.10∗ 0.24+ 0.11 0.03 0.14∗ considerably higher than that reported in the present ex- + Zn periment, even for the treatments with no micronutrient CV (%) 5.95 5.03 6.41 13.35 49.26 12.54 4.14 supply. WB, WCu, and WZn: control treatments, without application of B, Cu, and Lima et al. [63] determined the electrical conductivity of Zn, respectively; FS: foliar spray with boric acid, copper sulphate, and zinc beans subjected to B doses and verified that both the absence sulphate (0.4%); B: trunk injection of tablets containing B salts; Cu: trunk injection of tablets containing Cu salts; Zn: trunk injection of tablets and toxic concentrations of the nutrient are harmful to the containing Zn salts; B + Cu: trunk injection of tablets containing B and Cu quality of bean seeds, enhancing the electrical conductivity. salts; B + Zn: trunk injection of tablets containing B and Zn salts; Cu + Zn: Moreover, high physiological quality of bean seeds was trunk injection of tablets containing Cu and Zn salts; B + Cu + Zn: trunk obtained in a consortium with beans and castor beans, when injection of tablets containing B, Cu, and Zn salts; ∗mean values are sta- supplied with adequate doses of B. tistically different from those of the control treatment (FS) at the 10% significance level, according to Dunnett’s test; +mean values are statistically Evaluating the physiological quality of bean seeds over different from those of the control treatment (WB-WCu-WZn) at the 10% different doses of Mn and Zn, Teixeira et al. [64] did not significance level, according to Dunnett’s test. observe the effect of Zn on the electrical conductivity; however, adequate doses of Mn improve seed quality resulting in low electrical conductivity (65.8 μScm−1·g−1). Barrios [68] which was between 5.73 and 5.88. According to +e seed quality was the lowest in the control treatment, Sivetz and Desrosier [69], roasted beans of palatable coffees, without application of Zn and Mn. Amorim [16], Prete [35], without bitter or acidity, must have pH between 4.95 and 5.2. and Lima et al. [59] observed inverse correlation between +e results of the present work are slightly above the range leached potassium, electrical conductivity, and cupping established by this author. quality of coffee. Neves et al. [57] studying the effect of different doses of According to Amorim [65] and Chagas et al. [66], good Zn supplied to coffee plants by trunk injections of tablets coffees have high contents of sugars, around 8% according to containing Zn salts observed that the electrical conductiv- Navellier [67] and around 5 to 10% according to Prete [35]. ities of the control treatment, without Zn application, were In general, the results of the first crop season were above the 22.61 and 88.42 μS·cm−1·g−1 in two consecutive crop seasons, mean values reported by the literature. respectively. For the treatments with Zn inserted into the It can also be noted that total acidity remained below trunk, the means of the electrical conductivities were 16.84 211.2 g NaOH 100 mL−1 of the sample, considered by and 66.16 μS·cm−1·g−1. +e average values of leached po- Carvalho et al. [2] as a parameter for good coffees. With tassium were 1.13 and 0.95 g·kg−1, in two consecutive crop repect to the pH, it can also be observed that the mean of seasons, for the treatments without Zn application and 0.83 both years of assessment is quite close to that found by and 0.65 g·kg−1 for the treatments with Zn application. Journal of Food Quality 9

+e difference among the treatments was attributed to the +e contents of trigonelline observed in this work, in Zn functions on the cell membrane integrity of coffee beans. general, are in agreement with those established by the Correaˆ et al. [57] have also not observed variations in literature that vary from 0.6 to 1.2% for Coffea arabica [71], acidity of beans harvested from plants that received Zn by and the caffeine contents are close to those determined by injection of tablets in the trunk, with the average values Screenath [72] of about 1.2% for Coffea arabica. being 156.3 mL of NaOH 100 g−1, in coffees that were Within the monosaccharides and oligosaccharides, su- classified as “hard” in the cup test. In the same orchard, crose is a nonreducing sugar in greater quantity in coffee Martinez et al. [56] did not observe the effect of the nutrient beans, varying from 1.9 to 10% of the dry matter [73, 74]. on the total titratable acidity and pH, with the average values According to Knopp et al. [75], sucrose represents more than being 14.7 mL of NaOH 100 g−1 and 5.4 mL of NaOH 90% of the total low molecular weight carbohydrates and 100 g−1, respectively. corresponds to 7.07% of the dry matter of coffee beans; this value is very close to that found in this experiment. According to Camacho-Cristobal et al. [76], glucose 6- 3.5. Caffeine, Trigonelline, Glucose, Galactose, Arabinose, and phosphate, an enzyme involved in the glycolysis route, in Mannose. In this study, the effect of B, Cu, and Zn on the conditions of B sufficiency appears complexed with borate caffeine, trigonelline, sucrose, and glucose productions was anion and thus restricts the flow of the respiratory substrate evident by the significant differences between the sprayed to the pentose phosphate pathway; therefore, when B is treatment and the treatments that received the nutrients via adequate, the sucrose production is greater. solid injections compared to the control treatments WB, Brown and Clark [77] reported that Cu-deficient wheat WCu, and WZn (Table 7). plants had significantly lower soluble carbohydrate contents It is possible to observe the effect of the different ways of than well-fertilized plants. +e lower levels of plastocyanin, B, Cu, and Zn supply on the contents of caffeine, trig- as a consequence of Cu deficiency, may decrease the effi- onelline, and sucrose by the difference between the treat- ciency of photosynthetic electron transport in photosystem I ments with B + Cu + Zn and the sprayed treatment (Table 7). and thus impair the CO2 fixation rate, in such a way that +e results suggest that the lack of B, Cu, and Zn influenced starch and soluble sugars content (especially sucrose) are the route of caffeine and trigonelline synthesis, but probably reduced. the excessive concentrations of B, like in the treatment Coffee bean contents of specific sugars such as mannose, containing B + Cu + Zn, also did. gallactose, glucose, and arabinose were slightly lower than +ere was no effect of B on galactose production. +e those previously reported by Fischer et al. [78] for Coffea arabinose of the treatments containing B + Cu and B + Zn arabica. presented means statistically higher than those presented by the control treatment (WB). Regarding mannose, only the treatment containing B + Cu was statistically different from 3.6. Caffeoylquinic Acids (3-CQA, 4-CQA, and 5-CQA), PPO the control treatment WB, suggesting the effect of B on its Activity, and Phenolic Compounds. Among the phenolic production (Table 7). compounds, caffeoylquinic acids, dicaffeoylquinic acids, and +e effect of Cu on mannose levels in the coffee beans feruloylquinic acids are the main chlorogenic acid subgroups was evidenced by the significant difference among the beans present in coffee. In general, these compounds react during produced from the plants of the sprayed treatment and those the roasting process producing free phenolic acids and receiving tablets of Cu, Cu + B, and Cu + Zn inserted in the therefore volatile phenolic compounds that contribute to the trunk compared to the treatment WCu. Regarding the levels aroma of coffee beans [37]. of galactose, only the sprayed treatment differed from the +e effects of B, Cu, and Zn on 3-CQA and 5-CQA treatment WCu, and arabinose was significantly greater only contents were evidenced by the significant difference be- in the treatment containing Cu + B inserted in the trunk tween the treatments that received the nutrients via solid (Table 7). injections or foliar sprays and the treatments WB, WCu, and With regard to the galactose, significant differences were WZn. Compared to the sprayed treatment, only the treat- not observed between the treatments with Zn applications. +e ment with B + Cu + Zn differed significantly, for 5-CQA arabinose content in beans of the treatments containing Zn content, responding to the different ways of B, Cu, and Zn and Zn + B was statistically greater than that of the treatment supply (Table 8). WZn. Mannose of the sprayed treatment and of those re- +e 4-CQA, proanthocyanidin, and total phenolic ceiving Zn and Cu + Zn inserted in the trunk differed from compounds were not affected by B treatments. +e effect of that of the treatment WZn, showing the Zn effect on Cu, also, was not significant for the contents of 4-CQA and monosaccharides synthesis and the close relationship between proanthocyanidin. For Zn, 4-CQA and proanthocyanidin its production and the content of Zn in index leaves (Table 7). were not significant (Table 8). Mazzafera [70], working with nutritive solution and +e effect of B on PPO activity was evidenced by the young coffee plants, did not found significant effects of B, significant difference of the treatments with B + Cu and the Cu, and Zn deprivation on the caffeine production by coffee sprayed treatment compared to the treatment WB, even leaves. +e author states that the effect of mineral nutrients when B concentration in the index leaves of the treatment on the activity of methyltransferases involved in caffeine that received B by trunk injections was above the adequate synthesis is still unclear. range established by Martinez et al. [43]. 10 Journal of Food Quality

Table 8: 3-Caffeoylquinic acid (3-CQA, %), 4-caffeoylquinic acid concentration in the index leaves, which indicates adequate (4-CQA, %), 5-caffeoylquinic acid (5-CQA, %), proanthocyanidin nutrition (Tables 2 and 8). −1 −1 (Pro, %), polyphenol oxidase activity (PPO, U·min ·g ), and Regarding the Zn nutrition, the PPO activity, and contents phenolic compounds (TP, %) of coffee beans harvested from plants of total phenolic compounds, only the treatment with Zn + Cu submitted to the fertilization via solid salts injections in the trunk and the sprayed treatment differed from the treatment WZn, or foliar sprays with B, Cu, and Zn, in the crop season 2011/2012. with the highest activity being observed when phenolic Treatments 3-CQA 4-CQA 5-CQA Pro PPO TP compounds concentrations were low (Table 8). Boron For the PPO activity, compared with the sprayed WB 0.65∗ 0.73 1.69∗ 6.36 74.52∗ 6.37 treatment, all treatments that received B, Cu, and Zn via FS 0.44+ 0.72 1.41+ 5.70 85.93+ 5.15 solid injections, except B + Cu + Zn, are statistically similar, B 0.44+ 0.58 1.35+ 5.53 82.07 4.21 confirming the equal effect of different ways of supply of the + + + B + Cu 0.47 0.66 1.40 6.78 85.50 4.61 studied micronutrients (Table 8). B + Zn 0.45+ 0.64 1.38+ 7.14 78.38 5.50 ∗ ∗ Several works, in the literature, relate the accumulation B + Cu + Zn 0.52 0.61 1.58 6.41 74.68 6.13 of caffeoylquinic acids to B deficiency. Camacho-Cristobal CV (%) 19.89 22.27 6.36 38.15 9.01 24.59 Copper et al. [76] reported that the main effect of B deficiency is the WCu 0.65∗ 0.73 1.69∗ 6.36 74.53∗ 6.38 accumulation of glucose, fructose, and starch, followed by an FS 0.44+ 0.72 1.41+ 5.70 85.93+ 5.15 increase in the 3-CQA, 4-CQA, and 5-CQA contents in Cu 0.47+ 0.75 1.40+ 6.23 85.15+ 4.06+ tobacco leaves. +erefore, the high concentration of phe- B + Cu 0.47+ 0.66 1.39+ 6.78 85.51+ 4.61 nolic compounds, in B-deficient plants, could be a result of Cu + Zn 0.44+ 0.62 1.30+ 7.59 85.19+ 4.87 the soluble sugars accumulation [5]. B + Cu + Zn 0.52+ 0.61 1.58∗ 6.41 74.68∗ 6.14 Camacho-Cristobal et al. [79] attributed this effect to the CV (%) 16.10 14.37 5.45 37.40 6.25 22.57 enhancement of the phenylalanine ammonia lyase activity Zinc and consequent increase in the phenolic compounds syn- WZn thesis. Additionally, according to this author, when in high T 0.65∗ 0.73 1.69∗ 6.36 74.52∗ 6.37∗ concentration, B and glucose 6-phosphate form complexes FS 0.44+ 0.72 1.41++ 5.70 85.93+ 5.15+ Zn 0.48+ 0.74 1.41+ 7.21 79.08 5.70 and therefore restrict the flow of the respiratory substrate for B + Zn 0.45+ 0.64 1.38+ 7.14 78.38 5.50 the pentose phosphate pathway. Such a behavior may ex- Cu + Zn 0.44+ 0.61 1.30+ 7.59 85.19+ 4.87+ plain the enhanced concentrations of the 3-CQA and 5-CQA B + Cu + Zn 0.52+ 0.60 1.58∗ 6.41 74.68∗ 6.13 in the control treatment WB, WCu, and WZn. CV (%) 16.25 18.28 6.08 31.93 7.11 14.01 In addition, phenolic compounds accumulation is WB, WCu, and WZn: control treatments, without application of B, Cu, and a feature of B-deficient plants because of the formation of Zn, respectively; FS: foliar spray with boric acid, copper sulphate, and zinc borate complexes with some phenols that can be involved in sulphate (0.4%), B: trunk injection of tablets containing B salts, Cu: trunk the regulation of free phenol concentration and in the al- injection of tablets containing Cu salts, Zn: trunk injection of tablets cohol phenol synthesis, which are direct precursors of the containing Zn salts, B + Cu: trunk injection of tablets containing B and Cu salts, B + Zn: trunk injection of tablets containing B and Zn salts, Cu + Zn: lignin [80]. trunk injection of tablets containing Cu and Zn salts, B + Cu + Zn: trunk A strong relationship between caffeoylquinic acid con- injection of tablets containing B, Cu, and Zn salts; ∗mean values are sta- tents and cupping quality was not observed in the present tistically different from those of the control treatment (FS) at the 10% + study. Coffee beans from the control treatment (WB-WCu- significance level, according to Dunnett’s test; mean values are statistically WZn) had an average score of 82.7 and were hence classified different from those of the control treatment (WB-WCu-WZn) at the 10% significance level, according to Dunnett’s test. as “soft.”

3.7. PPO. Hajiboland and Farhanghi [81] studying the effect It is suggested that the content of B that is good for great of adequate and low doses of B in turnip plants observed that growth and production is below the content that maximizes PPO activity and phenolic compounds increased in roots the PPO activity, a feature that has been directly related to and shoot when the B supply was low. +e PPO activity in cupping quality (Tables 2 and 8). leaves and roots of deficient plants was 6.3 and 4.6 folds A significant difference was observed between the sprayed higher, respectively, than that of the control plants receiving treatment and the treatments with Cu, Cu + B, and Cu + Zn sufficient B. compared to the treatment WCu for the PPO activity, with According to Karabal et al. [82], excess of B alters the cell a particular focus on the clear inverse relationship between the membrane integrity; thus, a progressive increase in PPO PPO activity and the concentration of 5-CQA in the beans activity is observed initially followed by falling, because of (Table 8). quinones production that inhibits the enzyme, which may Even though average leaf Cu concentrations of sprayed explain the low PPO activity of the treatment containing treatments were somewhat lower than those of trunk in- B + Cu + Zn inserted into the trunk. jection treatments, they were still within the sufficiency Carvalho et al. [2] proposed a way to assess the coffee range reported in the literature of Martinez et al. [43]. It quality using PPO activity levels. According to them, “rio” suggests that the composition of caffeoylquinic acids and and “rioysh” types are well correlated to PPO activities below PPO activity remained constant within the range of Cu 55.99 U·min−1·g−1 of the sample; hard type is correlated to Journal of Food Quality 11 activities between 55.99 and 62.99 U·min−1·g−1 of the Copper and Zn supplied by solid trunk injections give sample; soft type is correlated to activities between 62.99 and equivalent results to foliar sprays, both in production and 67.66 U·min−1·g−1 of the sample, and strictly soft type is quality. correlated to activities above 67.99 U·min−1·g−1 of the Trunk injections of tablets containing B salts resulted in sample. toxicity and affected negatively the production, while some In the present work, it is not possible to establish a good attributes related to the quality of the grains were higher with relationship between cupping quality and PPO activity, since high B supply, capable of limiting the growth and yield of the treatments WB, WCu, WZn, and with B + Cu + Zn had coffee plants. high cupping test scores (82.7 to controls and 80.4 to combination: “soft” type) followed by low PPO activity. It Disclosure can be highlighted, however, that low PPO activity in these treatments was accompanied by high concentrations of 3- +is paper is part of the Ph.D. thesis presented to the CQA and 5-CQA. Universidade Federal de Viçosa, Viçosa, Brazil, by the first According to Mazzafera and Robinson [83], the 5-CQA author. is likely the main substrate of the PPO. Farah et al. [23] stated that there is an increase in the coloration intensity of coffee Conflicts of Interest beans in response to the PPO action on 5-CQA; thus, the +e authors declare that they have no conflicts of interest. authors associated the oxidation products with the low quality of coffee beans rich in caffeoylquinic acid. Amorim Acknowledgments et al. [84] reported that PPO action, along the structural changes of the membrane, is a possible cause for the for- +e authors would like to thank the financial support of the mation of beans classified as “rioysh” type. Brazilian government agencies Conselho Nacional de It is often mentioned by the literature that Cu is the PPO Desenvolvimento Cient´ıfico e Tecnologico´ (CNPq) and catalyst [85]; thus, the evaluation of PPO activity can be Coordenação de Aperfeiçoamento de Pessoal de N´ıvel Su- a good indicator of the nutritional status in Cu. Taking into perior (CAPES). account the importance of Cu on the PPO structure and because of the enzyme involved in the control of the con- References centration of free phenolic compounds, the efficiency in Cu supplementation to the plants is, therefore, a requirement of [1] E. Cantergiani, H. Brevard, Y. Krebs, A. Feria-Morales, great importance in order to obtain coffees with higher R. Amado,` and C. Yeretzian, “Characterization of mouldy/ quality. earthy defect in green Mexican coffee,” in Colloquium of In a nutritive solution experiment [86], it was observed International Coffee Science Association, Vol. 18, ASIC, Helsinki, Finland, 1999. that Zn doses affected the contents of chlorogenic acids of [2] V. D. Carvalho, S. M. Calfoun, S. J. R. Chagas, N. Botrel, and Coffea arabica beans. Total phenolic compounds, 5-CQA, E. S. G. Juste Júnior,“Relações entre a composição f´ısico- and 4-CQA reached minimum points when the index leaves qu´ımica e qu´ımica do grão beneficiado e da qualidade de −1 presented 10 mg·kg of Zn, that is about in the center of the bebida do café.Pesquisa AgropecuáriaBrasileira,” Bras´ılia, sufficiency range established by Martinez et al. [43]. +e vol. 29, no. 3, pp. 449–454, 1994. grains produced in conditions of deficiency or excess of Zn [3] E. Malavolta, “Nutrição, adubação e calagem para o cafeeiro,” presented higher values of these compounds. +e curves for in Cultura do Cafeeiro: Fatores que Afetam a Produtividade, PPO and 3-CQA presented exactly an inverse shape, A. B. Rena, E. Malavolta, M. Rocha, and T. Yamada, Eds., reaching the maximum points in grains of plants with pp. 136–274, Potafos, Piracicaba, Brazil, 1986. 10 mg·kg−1 of Zn in index leaves. Due to the direct re- [4] R. M. Prado, Nutrição de Plantas, UNESP, Jaboticabal, Brazil, lationship between 3-CQA and PPO, the author questioned 2008. [5] H. Marschner, Mineral Nutrition of Higher Plants, Academic if the content of 3-CQA, in a different way from that of 5- Press, New York, NY, USA, 2011. CQA, could be related to good quality of coffee beans. [6] J. R. Sanders, T. M. Adams, and B. T. Christensen, “Ex- Although, in this work, there was no good agreement tractability and bioavailability of zinc, nickel, cadmium and between the cupping test and chemical attributes of coffee copper in three Danish soils sampled 5 years after application quality, it should be emphasized that the cupping test is of sewage sludge,” Journal of Science Food and Agriculture, subjective and new methods must be studied in order to vol. 37, pp. 1155–1164, 1986. evaluate properly the coffee bean quality. [7] N. K. Fageria, V. C. Baligar, and R. B. Clark, Micronutrients in Crop Production, Vol. 77, Advances in Agronomy, New York, NY, USA, 2002. 4. Conclusions [8] E. Epstein and A. J. Bloom, Nutrição Mineral de Plantas: Princ´ıpiose Perspectivas, Planta, Londrina, Brazil, 2nd edition, Boron, copper, and zinc supplied by foliar sprays or solid 2006. trunk injections influence the chemical composition and [9] K. Mengel and E. A. Kirkby, Principles of Plant Nutrition, quality of the coffee beans, characterized by the contents of Internacional Potash Institute, Horgen, Switzerland, 4th caffeine, trigonelline, sucrose, glucose, arabinose, mannose, edition, 1987. 3-caffeoylquinic acid, 5-caffeoylquinic acid, polyphenol [10] R. Hansch and R. R. Mendel, “Physiological functions of oxidase activity, and total phenolic compounds. mineral micronutrients (Cu, Zn, Mn, Fe, Ni, Mo, B, Cl),” 12 Journal of Food Quality

Current Opinion in Plant Biology, vol. 12, no. 3, pp. 259–266, [27] E. Malavolta, G. C. Vitti, and S. A. Oliveira, Avaliação do 2009. Estado Nutricional das Plantas: Princ´ıpios e Aplicações, [11] I. Cakmak, H. Marschner, and F. Bangerth, “Effect of zinc Potafos, Piracicaba, Brazil, 2nd edition, 1997. nutritional status on growth, protein metabolism and levels of [28] Association of Official Analytical Chemists, Official Methods of indole-3 acetic acid and other phytohormones in bean Analysis, AOAC, Washington, DC, USA, 12th edition, 1975. (Phaseolus vulgaris L.),” Journal Experimental Botany, vol. 40, [29] C. M. Johnson and A. Ulrich, Analytical Methods for Use in Plants pp. 405–412, 1989. Analyses, University of California, Los Angeles, CA, USA, 1959. [12] G. S. Valladares, G. C. G. Santos, C. A. Abreu, O. A. Camargo, [30] N. S. Sakyiama, H. E. P. Martinez, M. A. Tomaz, and A. Borém, and J. P. Ferrero, Zinco Total e Dispon´ıvel em Amostras de Cafe´ Ara´bica: do Plantio a` Colheita, UFV, Viçosa, MG, Brazil, Perfis de Solos do Estado de São Paulo, Bragantia, vol. 68, no. 4, 2015. pp. 1105–1114, Bragantia, Campinas, Brazil, 2009. [31] Association of Official Analytical Chemists, Official Methods [13] R. H. Santori, A. E. Boaretto, F. C. A. Villanueva, and of Analyses of the Association of Official Analytical Chemists, H. M. G. Fernandes, “Absorção radicular e foliar de 65Zn e sua AOAC, Washington, DC, USA, 15th edition, 1975. redistribuição em laranjeiras,” Revista Brasileira de Fruti- [32] N. Nelson, “A photometric adaptation of Somogy method for cultura, vol. 30, no. 2, pp. 523–527, 2008. the determination of glucose,” Journal of Biological Chemists, [14] N. A. V. D. Pinto, S. M. R. Fernandes, R. G. F. A. Giranda, and vol. 153, no. 1, pp. 370–380, 1944. V. D. Carvalho, “Avaliação de componentes qu´ımicos de [33] V. L. Singleton, “+e total phenolic content of grape berries padrões de bebida para preparo do cafe´ expresso. Cienciaˆ e during the maturation of several varieties,” American Journal Agrotecnologia,” Lavras, vol. 26, no. 4, pp. 826–829, 2002. of Enology and Viticulture, vol. 17, no. 2, pp. 126–134, 1966. [15] P. F. P. Goulart, J. D. Alves, E. M. Castro, D. D. Fries, [34] T. M. Loeffler, D. M. Tekrony, and D. B. Egli, “+e bulk M. M. Magalhães, and H. C. Melo, “Histological and mor- conductivity test as an indicator of soybean seed quality,” phological aspects of different grain coffee qualities,” Cienciaˆ Journal of Seed Technology, vol. 12, no. 1, pp. 37–53, 1988. Rural, vol. 37, pp. 662–666, 2007. [35] C. E. C. Prete, “Condutividade eletrica´ do exsudado de grãos [16] V. H. Amorim, Aspectos Bioqu´ımicos do Grão de Cafe´ Verde de cafe´ (Coffea arabica L.). Desenvolvimento da metodolo- Relacionados Com a Deterioração da Qualidade, Tese de Livre gia,” CiênciasAgrárias, vol. 21, no. 1, pp. 67–70, 1995. Docencia,ˆ p. 85, ESALQ-USP, Piracicaba, Brazil, 1978. [36] J. D. Ponting and M. A. Joslyn, “Ascorbic acid oxidation and [17] S. J. Chagas, V. D. Carvalho, and L. Costa, “Caracterização browning in apple tissue extracts,” Archives of Biochemistry, qu´ımica e qualitativa de cafes´ de alguns munic´ıpios de tresˆ vol. 19, pp. 47–63, 1948. regiões produtoras de Minas Gerais,” Pesquisa Agropecuária [37] A. Farah, T. De Paulis, L. C. Trugo, and P. R. Martin, “Effect of Brasileira, vol. 31, pp. 555–561, 1996. roasting on the formation of chlorogenic acid lactones in [18] L. C. Trugo and R. A. Macrae, “Chlorogenic acid composition coffee,” Journal of Agricultural and Food Chemistry, vol. 53, of instant coffee,” Analyst, vol. 109, pp. 263–266, 1984. no. 5, pp. 1505–1513, 2005. [19] M. N. Clifford and K. C. Wilson, Coffee: Botany, Biochemistry [38] P. Mazzafera, A. Crozier, and G. Sandberg, “Studies on the and Production of Beans Beverage, Croom Helm, New York, metabolic control of caffeine turnover in developing endo- NY, USA, 1985. sperms and leaves of Coffea arabica and Coffea dewevrei,” [20] C. P. Bicchi, O. M. Panero, G. M. Pellegrino, and A. C. Vanni, Journal of Agricultural and Food Chemistry, vol. 42, no. 7, “Characterization of green and roasted coffees through the pp. 1423–1428, 1994. chlorogenic acid fraction by HPLC-UV and principal com- [39] M. D. Vitorino, A. S. França, L. S. Oliveira, and ponent analysis,” Journal of Agricultural and Food Chemistry, M. L. A. Borges, “Metodologias de obtenção de extrato de café vol. 43, no. 6, pp. 1549–1555, 1995. visando a dosagem de compostos não volateis,”´ Revista [21] M. Nogueira and L. C. Trugo, “Distribuição de isômerosde Brasileira de Armazenamento, vol. 26, no. 3, pp. 17–24, 2001. acido´ clorogenicoˆ e teores de cafe´ına e trigonelina em cafes´ [40] A. E. Hagerman, Y. Zhao, S. E. Johnson, and F. Shahadi, solúveis brasileiros,” Ciência e Tecnologia de Alimentos, “Methods for determination of condensed and hydrolyzable vol. 23, no. 2, pp. 296–299, 2003. tannins,” ACS Symposium Series, vol. 662, pp. 209–222, 1997. [22] H. Morais, P. H. Caramori, M. S. Koguishi, J. C. Gomes, and [41] A. Sluiter, B. Hames, R. Ruiz et al., Determination of Structural A. M. Ribeiro, “Sombreamento de cafeeiros durante o desen- Carbohydrates and Lignin in Biomass. Laboratory Analytical volvimento das gemas florais e seus efeitos sobre a frutificação e Procedure, General Books LLC, Memphis, Tennessee, 2008. produção,” Cieˆncia Rural, vol. 39, pp. 400–406, 2009. [42] SAEG-System of Statistical Analysis and Genetics, Versão 9.1: [23] A. Farah, M. C. Monteiro, V. Calado, A. S. Franca, and Fundação Arthur Bernardes, Viçosa, Brazil, 2007. E. L. C. Trugo, “Correlation between cup quality and chemical [43] H. E. P. Martinez, J. F. S. Menezes, R. B. Souza, attributes of Brazilian coffee,” Food Chemistry, vol. 98, no. 2, V. H. Alvarez-Venegas, and P. T. G. Guimarães, “Faixas pp. 373–380, 2006. cr´ıticas de concentração de nutrientes e avaliação do estado [24] M. C. Monteiro and L. C. Trugo, “Determinação de com- nutricional de cafeeiros em quatro regiões de Minas Gerais,” postos bioativos em amostras comerciais de cafe´ torrado,” Pesquisa Agropecua´ria Brasileira, vol. 38, no. 6, pp. 703–713, Qu´ımica Nova, vol. 28, pp. 637–641, 2005. 2003. [25] E. B. Fagan, C. H. E. Souza, N. M. B. Pereira, and [44] P. H. Brown and B. J. Shelp, “Boron mobility in plants,” Plant V. J. Machado, “Efeito do tempo de formação do grão de cafe´ soil, vol. 193, pp. 85–101, 1997. (Coffea sp.) na qualidade da bebida,” Bioscience Journal, [45] P. H. Brown and H. Hu, “Phloem boron mobility in diverse vol. 27, pp. 729–738, 2011. plant species,” Botany Acta, vol. 111, pp. 331–335, 1998. [26] P. T. G. Guimarães, A. W. R. Garcia, V. H. Alvarez et al., [46] F. Santinato, F. L. Castanehira, R. M. Prado, and R. Santinato, “Cafeeiro,” in Recomendações Para o Uso de Corretivos e “Efeito da aplicação de doses elevadas de acido´ borico´ em Fertilizantes em Minas Gerais-5a Aproximação, A. C. Ribeiro, cafeeiro jovem, Catua´ı Vermelho IAC 144,” in Congresso P. T. G. Guimarães, and V. H. Alvarez, Eds., pp. 289–302, Brasileiro de Pesquisas Cafeeiras, vol. 38, pp. 243-244, Procafe,´ Viçosa, MG, Brazil, 1999. Caxambu, RJ, Brazil, 2012. Journal of Food Quality 13

[47] O. F. Lima Filho and E. Malavolta, “Calibração de boro e zinco [64] I. R. Teixeira, A. Borem,´ G. A. A. Araújo,and M. J. B. Andrade, para o cafeeiro (Coffea arabica L. cv. Catua´ı Amarelo),” in Teores de Nutrientes e Qualidade Fisiológicade Sementes de Reunião Brasileira de Fertilidade do Solo e Nutrição de Plantas, Feijão em Resposta a` Adubação Foliar com Manganesˆ e Zinco, vol. 20, pp. 54-55, Potafos, Piracicaba, Brazil, 1992. vol. 64, no. 1, pp. 83–88, Bragantia, Campinas, Brazil, 2005. [48] A. B. Rena and M. Maestri, “Fisiologia do cafeeiro,” Informe [65] H. V. Amorim, M. G. Legendre, V. L. Amorim, A. J. S. Angelo, Agropecuário, vol. 11, pp. 26–40, 1985. and R. L. Ory, “Chemistry of Brazilian green coffee and the [49] R. Santinato, R. P. Camargo, R. B. Batista, and E. M. Pereira, quality of the beverage. VII. Total carbanyls, activity of “Quantificação de boro orgânico(ager boro) via foliar para o polyphenol oxidase, and hydroperoxides,” Turrialba, vol. 26, cafeeiro, seu efeito sobre aplicação na florada e seu com- no. 2, pp. 193–195, 1976. portamento no solo,” in Congresso Brasileiro De Pesquisas [66] S. J. R. Chagas, V. D. Carvalho, L. Costa, and Cafeeiras, vol. 20, A´ guas de Lindoía, Brazil, 1994. M. M. Romaniello, “Caracterização qu´ımica e qualitativa [50] U. V. Barros, R. Santinato, and J. B. Matiello, “Fontes, doses e de cafes´ de alguns munic´ıpios de tresˆ regiões produtoras de metodo´ de aplicação de B no plantio de cafe´ na região da Zona Minas Gerais. II-valores de acidez titulavel´ total e teores de da Mata em Minas Gerais,” in Congresso Brasileiro De Pes- açúcares (redutores, não redutores e totais),” Cienciaˆ e quisas Cafeeiras, vol. 22, pp. 117-118, Instituto Brasileiro do Agrotecnologia, vol. 20, no. 2, pp. 224–231, 1996. Cafe,´ Aguas´ de Lindoia,´ Brazil, 1996. [67] P. Navellier, “Coffee,” in Encyclopedia of Industrial Chemical [51] V. M. M. Andrade, Influenciaˆ do Cobre no Crescimento, Analysis, vol. 19, pp. 73–447, John Wiley & Sons, New York, Morfologia e Anatomia da Folha e na Composição Mineral do NY, USA, 1970. Cafeeiro, Tese (Doutorado), Universidade Estadual de São [68] B. B. E. Barrios, Caracterização F´ısica, Qu´ımica, Micro- Paulo, Jaboticabal, Brazil, 1973. biologica´ e Sensorial de Cafes´ (Coffea arabica´ L.) da Região Alto [52] J. R. Loneragan, “+e availability and absorption of trace Rio Grande-Sul de Minas Gerais, Dissertações (Mestrado em elements in soil-plants systems and their relation to move- Cienciaˆ dos Alimentos), p. 72, Universidade Federal de ment and concentrations of trace elements in plant,” in Trace Lavras, Lavras, 2001. Elements in Soil-Plant-Animal Systems, D. J. D. Nicholas and [69] M. Sivetz and N. W. Desrosier, Coffee Technology, Avi, A. R. Egan, Eds., pp. 109–134, Academic Press, Cambridge, Westport, CT, USA, 1979. London, 1975. [70] P. Mazzafera, “Mineral nutrition and caffeine content in [53] W. Reuther and C. K. Labanauskas, “Copper,” in Diagnosis coffee leaves,” Bragantia, vol. 58, no. 2, pp. 387–391, 1999. criteria for plants and soils, H. C. Chapman, Ed., pp. 157–179, [71] A. Illy and R. Viani, Espresso Coffee: Se Chemistry of Quality, University of California, CA, USA, 1966. Academic Press, San Diego, CA, USA, 1995. [54] P. T. G. Guimarães, J. C. Carvalho, C. C. A. Melles, and [72] H. L. Screenath, Development of Caffeine in Free Coffee Va- E. Malavolta, “Efeitos da aplicação foliar de sulfato de Zn na rieties, vol. 61, no. 10, Indian Coffee, Bangalore, India, 1997. produção e na composição mineral das folhas de cafeeiro [73] M. L. Wolfrom, R. A. Plunkert, and M. L. Laver, “Carbo- (Coffea arabica L.),” Anais da ESALQ, vol. 40, pp. 497–507, hydrates of the coffee bean,” Journal of Agricultural and Food 1983. Chemistry, vol. 8, no. 1, pp. 58–65, 1960. [55] O. F. Lima Filho and E. Malavolta, “Evaluation of extraction [74] J. R. Feldman, W. S. Ryder, and J. T. Kung, “Importance of procedures on determination of critical soil and foliar levels of non-volatile compounds to the flavor of coffee,” Journal of boron and zinc in coffee plants,” Commununications in Soil Agriculture and Food Chemistry, vol. 17, no. 4, pp. 733–739, Science and Plant Analysis, vol. 29, pp. 825–833, 1998. 1969. [56] H. E. P. Martinez, Y. Poltronieri, A. Farah, and D. Perrone, [75] S. Knopp, G. Bytof, and D. Selmar, “Influence of processing on “Zinc supplementation, production and quality of coffee the content of sugars in green Arabica coffee beans,” European beans,” Revista Ceres, vol. 60, no. 2, pp. 293–299, 2013. Food Research and Technology, vol. 223, no. 2, pp. 195–201, [57] Y. P. Neves, H. E. P. Martinez, and P. R. Cecon, “Effect of zinc 2006. and its form of supply on production and quality of coffee [76] J. J. Camacho-Cristobal, L. Lunar, F. Lafont, A. Baumert, and beans,” Journal of Science Food and Agricultural, vol. 91, A. Gonzales-Fontes, “Boron deficiency causes accumulation pp. 2431–2436, 2011. of chlorogenic acid and caffeoyl polyamine conjugates in [58] P. C. Correa,ˆ P. C. Afonso Junior, F. A. C. Pinto, and tobacco leaves,” Journal of Plant Physiology, vol. 161, no. 7, T. T. Oliveira, “Efeito da temperatura de secagem na cor dos pp. 879–881, 2004. grãos de cafe´ pre-processados´ por via seca e via úmida,” [77] J. C. Brown and R. B. Clark, “Copper as essential to wheat Revista Brasileira de Armazenamento, vol. 5, pp. 22–27, 2002. reproduction,” Plant Soil, vol. 48, no. 2, pp. 509–523, 1977. [59] M. V. Lima, H. D. Vieira, M. L. L. Martins, and [78] M. Fischer, S. Reimann, V. Trovato, and R. J. Redgwell, S. M. F. Pereira, “Preparo do cafe´ despolpado, cereja des- “Polysaccharides of green Arabica and Robusta coffee beans,” cascado e natural na região sudoeste da Bahia,” Revista Ceres, Carbohydrate Research, vol. 330, no. 1, pp. 93–101, 2001. vol. 55, pp. 124–130, 2008. [79] J. J. Camacho-Cristobal, D. Anzellotti, and A. Gonzalez- [60] W. Heydecker Vigour, in Viability on Seeds, E. D. Roberts, Fontes, “Changes in phenolic metabolism of tobacco plants Ed., pp. 209–252, University Press, Syracuse, NY, USA, 1972. during short-term boron deficiency,” Plant Physiology Bio- [61] J. F. Harrington, “Problems of seed storage,”in Seed Ecology, chemistry, vol. 40, no. 12, pp. 997–1002, 2002. W. Heydecker, Ed., pp. 251–263, +e Pennsylvania State [80] D. J. Pilbeam and E. A. Kirkby, “+e physiological role of University Press, University Park, PA, USA, 1973. boron in plants,” Journal of Plant Nutrition, vol. 6, no. 7, [62] M. R. Malta, M. L. Santos, and F. A. M. Silva, “Qualidade de pp. 363–382, 1983. grãos de diferentes cultivares de cafeeiro (Coffea arabica L.),” [81] R. Hajiboland and F. Farhanghi, “Remobilization of boron, Acta Scientiarum, vol. 24, no. 5, pp. 1385–1390, 2002. photosynthesis, phenolic metabolism and anti-oxidant de- [63] A. D. Lima, T. V. A. Viana, B. M. Azevedo, A. B. Marinho, and fense capacity in boron-deficient turnip (Brassica rapa L.) J. M. L. Duarte, “Adubação boracica´ na cultura do girassol,” plants,” Soil Science Plant Nutrition, vol. 56, no. 3, pp. 427– Revista Agro@mbiente, vol. 7, no. 3, pp. 269–276, 2013. 437, 2010. 14 Journal of Food Quality

[82] E. Karabal, M. Yücel,and H. A. Oktem,¨ “Antioxidant re- sponses of tolerant and sensitive barley cultivars to boron toxicity,” Plant Science, vol. 164, no. 6, pp. 925–933, 2003. [83] P. Mazzafera and S. P. Robinson, “Characterization of polyphenol oxidase in coffee,” Phytochemistry, vol. 55, pp. 285–296, 2000. [84] H. V. Amorim, A. R. Cruz, S. T. Angelo, A. J. Dias, R. M. Mello, and A. A. Teixeira, “Biochemical, physical and organoleptical changes during raw coffee quality deterioration,” in Proceedings of 8th International Colloq Coffee, Abidjan, Coteˆ d’Ivoire, December 1977. [85] E. Malavolta, Elementos de Nutrição Mineral de Plantas, Agronomicaˆ Ceres, São Paulo, Brazil, 1980. [86] J. S. Lacerda, Produção, Composição Qu´ımica e Qualidade da Bebida de Cafe´ Arabica´ em Razão da Dose de Cobre e Zinco, Tese de Doutorado, Universidade Federal de Viçosa, Viçosa, Brazil, 2014. [87] J. M. Clemente, Boron, Copper and Zinc Effects on Photo- synthesis, Enzymatic Activity, Nutritional Status, Production, Chemical Composition and Cup Quality of Coffee, Tese (Doutorado em Fitotecnia), p. 115, Universidade Federal de Viçosa, Viçosa, MG, Brazil, 2014. Hindawi Journal of Food Quality Volume 2018, Article ID 5908463, 7 pages https://doi.org/10.1155/2018/5908463

Research Article Quality of Commercial Coffees: Heavy Metal and Ash Contents

Mariana Teixeira Pigozzi, Flávia Regina Passos , and Fabrícia Queiroz Mendes

Institute of Agricultural Sciences, Federal University of Vic¸osa, Rio Parana´ıba Campus, Road MG 230, Km 07, P.O. Box 22, 38810-000 Rio Parana´ıba, MG, Brazil Correspondence should be addressed to Flavia´ Regina Passos; [email protected]

Received 29 September 2017; Accepted 28 March 2018; Published 6 May 2018

Academic Editor: Gabriel Henrique Horta de Oliveira

Copyright © 2018 Mariana Teixeira Pigozzi et al. Tis is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Tis study aimed to quantify the ash content and to determine the concentration of heavy metals in roasted ground cofee and their respective infusions. Te ash content was determined by incineration of the samples, and the quantifcation of heavy metals was performed by fame atomic absorption spectrophotometry for the following metals: cadmium, lead, copper, chromium, nickel, and zinc. According to the ash analysis, 15% of the roasted ground cofee samples were within the standards established by the legislation of the state of Sao˜ Paulo, which has set an ash content of below 5%. In the roasted ground cofee samples, the Cd, Cu, Ni, and Zn contents did not exceed the limit established by Brazilian legislation. In several samples, both Pb and Cr were found in high levels, exceeding the limits established by Brazilian legislation. In the infusions of roasted ground cofee, the Cd, Cu, Cr, and Ni contents were below the detection limit of the equipment. Zn was found in all infusions and Pb was only detected in seven cofee infusion samples. Overall, the concentrations of heavy metals found in the commercially roasted ground cofee and their respective infusions are lower than the limits recommended by the ofcial inspection agencies and, thus, are suitable for consumption.

1. Introduction price factor, which demonstrates consumer demand for the best cofee [4], free of impurities and adulterants, and is a Cofeebeansareoneofthemostwidelytradedcommodities genuine product. in the world. Tere are more than 100 diferent plant species Normative Instruction No. 16 of the Ministry of Agricul- of the Cofea genusintheworld.Despitethisgreatdiversity, ture and Livestock (MAPA), which guarantees the quality of only two species have signifcant economic importance in the roasted ground cofee, establishes an acceptable level of up world cofee market: Cofea arabica L. and Cofea canephora to 1% of impurities, sediments, and foreign matter in cofee Pierre ex A. Froeher, known, respectively, as arabica cofee [5]. However, there are other compounds that, when present and robusta cofee or conilon [1]. in the grains, could cause serious damage to the health of In the international market arabica cofee represents consumers [6]; this is the case for metal contaminants. more than 70% of world production. Tey are fner cofees Heavy metals are the most frequently evaluated elements recognized for the best aroma and favor, being more valued in food, because of their ability to accumulate in the food than conilon cofee. In addition, it is more commonly used in chain [6–11]. Tus, their maximum levels have become blends of arabica cofee, which gives more body and reduces quality standards around the world [10]. Tese elements are the acidity of the beverage [2]. stable and, thus, persistent in the environment, accumulating Roasted and ground cofee is classifed as Traditional or in the soil [8] as a result of factors associated with the weath- Extraforte (arabica blended with conilon, without quantity ering of rocks and soil formation, the soil soluble content, restriction) and Gourmet (arabica only), according to the pH, plant species, environmental conditions, technological characteristics of the products identifed through the cofee practices, and use of chemical products [6, 7, 12]. Among bean species, roasting, grinding, aroma, and favor [3]. Cofee thechemicalproductsthatcontributetoincreasedheavy consumers are becoming increasingly demanding. During metal content in the soil are biosolids, which ofen contain the purchasing process, quality factors and purity stamps are high levels of cadmium, copper, chromium, lead, nickel, and frequently sought characteristics. Tese factors outweigh the zinc [13]. Additional sources of contaminants include the use 2 Journal of Food Quality

Table 1: Detection limits (DL) and limits quantifcation (QL) determined for elements quantifed by atomic absorption.

−1 −1 −1 −1 −1 −1 Cu (mg kg )Cr(mgkg)Ni(mgkg)Cd(mgkg)Pb(mgkg)Zn(mgkg) DL 0.0474 0.3105 0.0414 0.1569 0.1104 0.7262 QL 0.1438 0.9408 0.1559 0.4775 0.3347 2.2007 of phosphate fertilizers produced from sedimentary rocks hollow cathode lamp (HCL) was used. Te intensities of the [14] and the application of micronutrients from industrial electric currents were 7 mA (Cr), 5 mA (Mn, Pb, and Zn), byproducts [7]. and 4 mA (Cd, Cu, and Ni). Te absorbances were measured First, these materials are absorbed by plants and accu- at the following wavelengths (nm): Cd, 228.8; Cu, 324.7; Cr, mulate in food, thus becoming sources of contamination 357.9; Pb, 217.0; Ni, 232.0; and Zn, 213.9. Te procedure used for humans [7]. Magna et al. [9] detected the presence of and the appropriate concentrations for the standard curve Cd and Pb in many types of food and grasses cultivated were performed according to the protocols established in the in the municipality of Santo Amaro, Bahia. Schmidt et al. equipment manual. [15] observed the presence of many metals in cofee grains Deionized water was used for all analyses. Te con- cultivated in the state of Parana,´ some at concentrations much centrations were determined by elaborating the calibration higher than the values established as safe for daily intake. analytical curves of each element quantifed. Te solu- Terefore, this study aims to determine, by means of tions for the construction of the standard curves were incineration, the ash content and the fame atomic absorption obtained through stock solutions of 1000 mg/L for atomic spectrophotometry, the heavy metal contents of commercial absorption and ICP. Te reagent blanks and the samples samples of roasted ground cofee, and their respective infu- with the standards of the elements were included in the sions. analysis. Te detection limit (DL) and the quantifcation limit 2. Material and Methods (QL) were calculated based on the standard deviation of the responseandtheslopeofthecalibrationcurve(meanof3 Te cofee samples were purchased in the local markets of the curves). Tey were obtained from the following equations cities of Rio Parana´ıba and Carmo do Parana´ıba in the state [17]: of Minas Gerais. Fifeen samples of commercially roasted ground cofee were analyzed, three of which were considered = 3.3SD , “Gourmet”(samplesI,N,andO).Twosamplesofeachcofee DL � were purchased to perform the analyses. Two replicates of (1) eachsampleweretakentotheanalysis. =10SD , QL � For the determination of the ash content in the roasted ground samples, approximately 2 g of sample was weighed where in a porcelain capsule and placed in a mufe furnace for ∘ incineration at 550 Cuntiltheorganicmatterhadbeenelim- SD is standard deviation of the response or instru- inated. At the end of the incineration process, the samples mental target (3 curves); were weighed. Te results are expressed in percentages [16]. � Te cofee infusion consisted of the beverage prepared is slope or angular coefcient of the analytical curve. with 12 g of roasted ground cofee lixiviated by 100 mL ∘ of water heated to boiling point (95–100 C), followed by Te limits of detection and quantifcation determined fltration [16]. Ten, 25 mL of the beverage was concentrated experimentally for each of the evaluated elements are ∘ in an oven at 60 C with air circulation until reaching a fnal described in Table 1. volume of approximately 2.5 mL. For optical imaging, the roasted ground cofee samples Te samples of roasted ground cofee and the infusions were placed on slides and humidifed. Ten, they were an- were mineralized by the wet route, using 18 mL of the alyzed using an optical microscope and digitalized (5-mega- nitric–perchloric digestive mixture in a 3 : 1 ratio (nitric pixel resolution) using a digital camera (Samsung, WB150F, ∘ acid : perchloric acid) and heating at 190 C. Te solution was Manaus, AM, Brazil), CoolSnap Pro image system, and Im- maintained under these conditions until a translucent solu- age-Pro Plus sofware (Media Cybernetics). tion had formed [6]. Te samples were transferred to 25 mL Te obtained data were analyzed by descriptive statistics. volumetric fasks, and the volume was made up with water. Te heavy metal data in the roasted ground cofee samples Ten, the contents of the following elements were quantifed: were submitted to principal component analysis. Ten, the cadmium,copper,chromium,lead,nickel,andzinc.Temea- samples were grouped by Euclidean distance, using the values surements were made in a fast, sequential atomic absorption of the frst and second main components. spectrophotometer(Varian,A240FS,Mulgrave,VIC,Aus- Te Pearson correlation analysis was performed between tralia) with atomization in an air/acetylene fame at the rate of the concentrations of Cd, Cu, Cr, Pb, Ni, and Zn in −1 −1 −1 −1 13.3 L min /2.9 L min forCrand13.5Lmin /2.0 L min the samples of roasted ground cofee and their respective for the other elements. As a radiation source, a single element infusions. Journal of Food Quality 3

10 of the present study. dos Santos et al. [26] found mean Cu −1 contents from 7.15 to 14.9 mg kg in cofees produced in two 8 properties in the state of Bahia. Tese results indicate that the mineral content is associated with the origin of the cofee, species, pesticides, and agricultural treatment, among other 6 factors. Te roasted ground cofee samples had low Ni concentra- 4 ASHES (%) tions. We could only quantify the Ni content in six samples, −1 andthesehadNicontentsrangingfrom0.037to0.72mgkg , −1 2 having a mean of 0.11 mg kg . Tese values do not exceed the maximum limit allowed by Brazilian legislation, which −1 0 establishes a maximum Ni content of 5 mg kg [18]. In the −1 ABCDEFGHI J KLMNO other samples, the Ni content was lower than 0.025 mg kg . Figure 1: Ash content of the roasted ground cofee samples. Morgano et al. [25] determined the mean Ni content of −1 4.76 mg kg , considering all raw cofee samples analyzed, −1 and a mean value of 1.21 mg kg , considering only the 3. Results and Discussion samples from the Alto Parana´ıba region, MG. −1 Te Zn contents varied between 3.31 and 25.97 mg kg , Figure 1 shows the mean ash content observed in the diferent −1 roasted ground cofee samples. Te ash content ranged from having a mean content of 8.07 mg kg . Te determined −1 values do not exceed the maximum Zn content allowed 4.48 to 9.44 g (100 g) ,wherethelowestmeanvaluewas −1 obtained in one of the “Gourmet” cofee samples (sample I). by Brazilian legislation of 50 mg kg [18]. Te mean Zn Teixeira et al. [16] analyzed diferent commercial cof- content found in the samples is similar to the values found by Morgano et al. [25], who determined a mean Zn content fee brands and found ash contents ranging from 3.99 to −1 −1 6.10 g (100 g) ,wherethelowestvaluewasalsofoundina of 8.33 mg kg in raw cofee. Mean contents higher than the results of the present study were found by dos Santos et al. “Gourmet” cofee sample. Gourmet cofee is composed only −1 of arabica cofee, which commonly has a lower ash content. [26], who obtained values from 25 to 45 mg kg of Zn when Te other cofees are composed of blends and the addition evaluating two cofee plantation properties in the state of Bahia. In addition, Ashu and Chandravanshi [12] found a of conilon cofee contributes to the increase of ash content −1 [20]. Normative Instruction No. 16 of MAPA [5] does not mean value of 19 mg kg of Zn in commercial roasted ground specify the maximum ash content in roasted ground cofee. cofee samples. However, Resolutions Nos. 19 and 31 of the Sao˜ Paulo State Cu, Ni, and Zn are essential elements for the development Board of Agriculture and Supply (Secretaria de Agricultura ofcofeeplantsandareappliedviathesoilortheleaves; e Abastecimento do Estado de Sao˜ Paulo, SAASP) defne as therefore, they could also be present in the grains. Copper is standard the maximum fxed content of mineral residue to be an active ingredient of some of the pesticides used in cofee 5.00% for both “Traditional” and “Gourmet” roasted ground cultures, in the form of copper hydroxide, copper oxychlo- cofee [21, 22]. Terefore, considering the Sao˜ Paulo state ride, copper sulfate, or copper ethylenediaminetetraacetate resolutions, two samples (O and I) meet these standards. (EDTA), which allows the absorption of this element by the According to a study performed by Durek de Conti et al. plant, explaining its presence in the grain [11]. Furthermore, [20], this variation could be explained by the diference in these elements could also be present in the soil, varying the varieties and cofee blends, where the mean ash content depending on their origin [6]. in arabica ranged from 2.5 to 4.5%, whereas that of robusta Crwasonlydetectedinoneroastedgroundcofeesample, −1 cofee was 4.64%. It is also possible to associate the mineral where the content was 3.27 mg kg ,avaluemuchhigherthan composition of the cofee with the nutritional state of the the maximum limit allowed by Brazilian legislation, which −1 cofee plantation and its location [23]. According to Muller¨ establishes a maximum content of 0.1 mg kg [18]. In the et al. [24], when the ash content exceeds 5.00%, there may be other samples, Cr was not detected, and all these contents −1 asignifcantquantityofimpuritiesinthesample. were below 0.025 mg kg . dos Santos et al. [26] did not fnd Te heavy metal contents found in the roasted ground Cr in cofee samples from the state of Bahia. cofee samples are presented in Figure 2. −1 Amongtheevaluatedelements,CdandPbareconsidered Te Cu contents ranged from 1.55 to 29.85 mg kg , −1 the most toxic inorganic contaminants [8]. According to having a mean content of 16.49 mg kg .Forallroasted Resolution RDC No. 42 of 2013, the maximum limit of Cd −1 ground cofee samples analyzed, the Cu concentrations were in roasted ground cofee is 0.1 mg kg and that for Pb is −1 −1 below the maximum limit (30 mg kg ) defned by Brazilian 0.5 mg kg [19]. legislation [18]. Morgano et al. [25] obtained mean Cu values Cd was not found in the analyzed samples. Hence, the −1 −1 higher than the results of the present study (29.86 mg kg ), Cd content for all samples was below 0.025 mg kg .dos considering all the samples analyzed by the authors. However, Santos et al. [26] found mean Cd contents between 0.70 and −1 whenanalyzingonlythesamplesfromtheAltoParana´ıba 0.75 mg kg in cofee of two diferent properties in the state −1 region, MG, the value found (14.17 mg kg )issimilartothat of Bahia. 4 Journal of Food Quality

35 50 − −1 Maximum limit: 50 mg kg 1 Maximum limit: 30 mg kg 45 30 40 25 35 ) ) −1 20 −1 30 25 15 20 Zn kg (mg Cu (mg kg (mg Cu 10 15 10 5 5 0 0 ABCDEFGHI J KLMNO ABCDEFGHI J KLMNO

4 5 −1 3.5 Maximum limit: 5 mg kg 4 3 ) ) −1 2.5 −1 3 2 1.5 2 Ni (mg kg (mg Ni Cr (mg kg Cr (mg 1 1 − 0.5 Maximum limit: 0,1 mg kg 1 0 0 ABCDEFGHI J KLMNO ABCDEFGHI J KLMNO

2.5

2 ) −1 1.5

Pb (mg kg Pb (mg 1 − Maximum limit: 0,5 mg kg 1 0.5

0 ABCDEFGHI J KLMNO Figure 2: Mean content of heavy metals in roasted ground cofee samples and maximum contents establish by Brazilian legislation [18, 19].

Pb was undetectable only in one sample. Te Pb con- and inhibiting hemoglobin synthesis, in addition to being a −1 tents ranged from 0.14 to 2.59 mg kg ,havingameanof potential carcinogen [27]. −1 0.69 mg kg . Among the analyzed samples, eight had Pb Intheanalysisofmaincomponents(Figure3),thefrst contents above the maximum limit allowed by Brazilian leg- main component explains 76.61% of the variation of the −1 contents of heavy metals of the analyzed samples, while the islation (0.5 mg kg ) [19]. Most of the samples in the present second main component explains 12.37%. study originate from local production (Alto Parana´ıba region, Te distance between the samples suggests the existence MG), which indicates that a large part of the cofee produced of a diference in the heavy metal content of the samples. By in this region contains Pb at high levels, indicating that, calculating the Euclidean distance, it is concluded that the most likely, this element is found in high concentration samples difer in three groups. in the soils of the Alto Parana´ıba region, MG [6]. Pb Sample A showed a positive correlation with the frst main can act on the neurological system, causing damage to the component and a great distance from the other samples. It central and peripheral nervous systems, interfering in the was the sample that presented a greater content of heavy conversion of vitamin D and the homeostasis of calcium, woods. During the acid digestion, precipitate was observed Journal of Food Quality 5

2 � −1 J Table 2: Mean contents ( g50mL ) of heavy metals present in a cup (50 mL) of cofee infusion. I G1 A H O Samples Zn Pb B0 A 10.952 ± 2.963 0.838 ± 0.218 −2 −1F K L 0M 1 2 3 4 5 6 7 B12.375± 1.401 0.129 ± 0.039 N −1 D Component 2 (12.37%) Component C E C 9.105 ± 0.1086 nd Component 1 (78.61%) D9.202± 1.575 0.322 ± 0.086 Figure 3: Analysis of major components of heavy metal content E 4.998 ± 3.237 nd in the roasted ground cofee samples. Distinguished groups across F7.584± 0.239 nd Euclidean distance. G7.116± 2.649 nd H9.626± 0.217 1.289 ± 0.211 in sample A. Te presence of precipitate in acidic solution I 10.430 ± 2.756 nd and the high content of heavy woods compared to the other J9.822± 3.322 2.835 ± 0.172 samples suggests the presence of inorganic contamination by K9.833± 2.901 1.160 ± 0.179 material that contains silica, which is an insoluble mineral in L9.452± 1.043 0.258 ± 0.038 acid. M7.736± 0.043 nd Te other samples were divided into two other groups, N 14.635 ± 2.727 nd and the samples of gourmet cofee analyzed (I, N, and O) O 6.486 ± 0.206 nd are distributed in both groups, showing that the fact that Mean 9.626 0.455 they are better classifed sensorially is not related to the content of heavy metals. Te variations in the contents of nd: not detected. heavy metals in the roasted ground cofee can be explained by the chemical composition of the soil and the management and cultivation of the plants and cofee [15]. Arabica and conilon cofee have many distinct characteristics, namely, content of metals in the roasted ground cofee and their of botanical, agronomic, and morphological nature. Tese respective infusions. In addition to the diferences in the species can be blended into the commercial cofee industry, blends and edaphoclimatic diferences in the production of which comes from a wide variety of geographical areas. Tese the grains, because these are commercial samples, the grain characteristics may also have contributed to variations in the size of the cofee powder could infuence the extraction ratio heavy metal content [28]. for these metals, making their extraction easier or more In the cofee infusions (50 mL), Cd, Cu, Cr, and Ni difcult (Figure 4). were not detected. Zn was detected in all samples, having a Te mean extraction of Pb in the “Traditional” and “Extra −1 maximum content of 15.472 �g50mL and a minimum of Strong” cofee infusions, which have smaller grain sizes −1 −1 4.998 �g50mL ,havingameanvalueof9.63�g50mL (Figure 4(a)), was 16.96%. For the cofees of higher quality, (Table 1). Noel¨ et al. [29] found the mean value for Zn named “Gourmet,” of larger grain size (Figure 4(b)), Pb was −1 foracofeebeverageof14.25�g50mL in the cofee not detected in their infusions, even when high Pb contents infusions, a value higher than the mean and similar to were found in the roasted ground cofee. themaximumvaluefoundinthepresentstudy.Pb,which Whencomparingthedataofthepresentstudywiththatof was present in almost all roasted ground cofee samples the literature, diferences are observed in the concentrations (Figure 2), was only detected in seven cofee infusion samples, found in some elements. Rocks are natural sources of all −1 having contents ranging from 0.129 to 2.835 �g (50 mL) chemical elements found on Earth [6]. Released by weather- (Table 2). ing, the elements in the rocks are made available in the soil −1 Te maximum acceptable intake of Pb is 25 �gkg body and then are carried to rivers and groundwater, interacting weight per week [9], which represents an intake of 250 �g with the ecosystem where humans live. However, the natural perdaybasedonanadultmaleweighing70kg.Hence,the occurrence of these elements is not equally distributed on maximum daily intake of four cups of cofee represents 0.21% the surface of the Earth and problems may arise when their to 4.54% of the maximum Pb intake. concentrations are too low (defciency) or too high (toxicity) Te infusion process used to prepare the cofee has a [6]. strong infuence on the concentration of metals. Te lower Metal contamination could also be caused by human metal concentration in the infusions is most likely due to the activities, such as waste from mining, the steel industry, release of all elements in the form of their simple ions with the cosmetics industry, or agricultural activities. Te con- the cofee matrix [30]. tamination that afects the agricultural areas currently rep- Te Pearson correlation coefcients were estimated for resents a signifcant problem, given that many pollutants the Pb and Zn concentrations in the samples of roasted play an essential role in economic activities, for example, ground cofee and their respective infusions. For Pb, the pesticides and fertilizers [6, 31], and many of these products Pearson correlation coefcient was 0.1204, and for Zn it was can remain in the soil or water, thus contaminating food 0.0649, indicating that there is no correlation between the [11]. 6 Journal of Food Quality

(a) (b)

Figure 4: Microscopy images of grains of the “Traditional” roasted ground cofee sample (a) and a “Gourmet” cofee sample (b).

4. Conclusions [6]S.A.Silva,F.Q.Mendes,M.R.Reisetal.,“Determinationof heavy metals in the roasted and ground cofee beans and brew,” Te roasted and milled cofees quantifed in the present study African Journal of Agricultural Research,vol.12,pp.221–228, present high levels of ash indicating that products available on 2017. the market can contain high amounts of impurities. [7] D. Dem´ırezen and A. Ahmet, “Heavy metal levels in vegetables Formostofthequantifedsamples,theconcentrationsof in turkey are within safe limits for Cu, Zn, Ni and exceeded for heavy metals found in the commercial roasted ground cofees Cd and Pb,” JournalofFoodQuality,vol.29,pp.252–265,2006. are below the limits recommended by the ofcial inspection [8] Z.-Y. Hseu, S.-W.Su, H.-Y. Lai, H.-Y. Guo, T.-C. Chen, and Z.-S. institutions and, thus, are suitable for consumption. Chen, “Remediation techniques and heavy metal uptake by dif- Some roasted ground cofee samples have heavy metal ferent rice varieties in metal-contaminated soils of Taiwan: New contentsabovetheallowedlevels.Parameterssuchasthe aspects for food safety regulation and sustainable agriculture,” species and the origin of the cofee beans can contribute to Soil Science & Plant Nutrition,vol.56,no.1,pp.31–52,2010. variations in the contents of heavy metals. [9]G.A.M.Magna,S.L.Machado,R.B.Portella,andM.R. Tere is no correlation between the Pb and Zn contents Carvalho, “Lead and cadmium detected in plant foods and intheroastedgroundcofeeandtheirrespectiveinfusions grasses in Santo Amaro, Bahia,” Qu´ımica Nova,vol.36,pp.989– and, consequently, the beverages can be considered uncon- 997, 2013. taminated by these metals. [10] J. Malik, J. Szakova, O. Drabek, J. Balik, and L. Kokoska, “Determination of certain micro and macroelements in plant stimulants and their infusions,” Food Chemistry, vol. 111, no. 2, Conflicts of Interest pp. 520–525, 2008. [11] V.Souza, R. F. B. Sousa, L. C. Ofemann, V.Vaz Alves, and A. C. Te authors declare that they have no conficts of interest. G. Alves Jr., “Alternatives for remediation and decontamination of soils from Brazil,” African Journal of Agricultural Research, vol. 9, pp. 3197–3204, 2014. References [12] R. Ashu and B. S. Chandravanshi, “Concentration levels of met- [1] L. R. Cagliani, G. Pellegrino, G. Giugno, and R. Consonni, als in commercially available ethiopian roasted cofee powders “Quantifcation of Cofea arabica and Cofea canephora var. and their infusions,” Bulletin of the Chemical Society of Ethiopia, robustainroastedandgroundcofeeblends,”Talanta,vol.106, vol. 25, no. 1, pp. 11–24, 2011. pp.169–173,2013. [13]R.S.M.P.Nascimento,G.S.Carvalho,L.P.Passos,andJ. J. Marques, “Lead and zinc leaching in soil treated with iron [2] M. Defernez, E. Wren, A. D. Watson et al., “Low-feld 1H NMR smelting residues,” Pesquisa Agropecuaria´ Tropical,vol.40,pp. spectroscopy for distinguishing between arabica and robusta 497–504, 2010. ground roast cofees,” Food Chemistry,vol.216,pp.106–113,2017. [14] C. A. Grant and S. C. Sheppard, “Fertilizer impacts on cadmium [3] ABIC – Brazilian Association of Cofee Industries, Recom- availability in agricultural soils and crops,” Human and Ecolog- mended Quality Standard and Good Practices for Manufactur- ical Risk Assessment,vol.14,pp.210–228,2008. ing Roasted Cofee Beans and Roasted and Ground Cofee, Rio [15] C. A. P. Schmidt, E.´ Miglioranza, G. Nagashima, and F. Grecco, de Janeiro, Brazil, 2017. “Heavy metals concentration in cofee grains produced in [4] J.M.Y.PerosaandL.H.F.Abreu,“Economicaspectsandoppor- farming under basalt and Caiuasandstonesoils,”´ Cienciaˆ Rural, tunities in the market of quality cofees,” Pesquisa Agropecuaria´ vol.39,no.5,pp.1590–1593,2009. Tropical,vol.39,pp.144–150,2009. [16] O.R.Teixeira,F.R.Passos,andF.Q.Mendes,“Physico-chemical [5] Brazila. Ministry of Agriculture, Livestock and Food Supply. and microscopic quality of 14 trademark roasted and ground Normative Instruction No 16, May 24, 2010. Technical regula- cofee,” Cofee Science,vol.11,no.3,pp.396–403,2016. tionforroastedcofeebeansandroastedandgroundcofee. [17] M. E. Scwartz and I. S. Krull, “Validation of chromatographic FederalOfcialGazetteofBrazil.May,2010.Section1,11. methods,” Pharmaceutical Technology, vol. 2, pp. 12–20, 1998. Journal of Food Quality 7

[18] Brazil. Decree No 55871, March 26, 1965. Modifes the decree No [31] M. C. S. De Moura, G. C. Moita, and J. M. Neto, “Analysis and 50.040, January 24, 1961, concerning regulatory standards for assessment of heavy metals in urban surface soils of Teresina, the use of food additives, as amended by Decree No 691, March Piau´ı State, Brazil: A study based on multivariate analysis,” 13, 1962. FederalOfcialGazetteofBrazil. Abril, 1965. Section 1, Comunicata Scientiae,vol.1,no.2,pp.120–127,2010. 3610. [19] Brazil. Sanitary Surveillance Agency. Resolution No 42, August 29, 2013. MERCOSUL Technical Regulation on Maximum Limits of Inorganic Contaminants in Foods. Federal Ofcial Gazette of Brazil. August, 2013. Section 1, 33. [20]M.C.M.D.DurekdeConti,C.S.GoodKitzberger,M.B.S. Dos Santos Scholz, and S. H. Helena Prudencio, “Physical and chemical characteristics of exotic and conventional roasted and ground cofees,” Boletim Centro de Pesquisa de Processamento de Alimentos,vol.31,no.1,pp.161–172,2013. [21] Brazil. Secretariat of Agriculture and Supply of Sao˜ Paulo State. Resolution No 31, June 22, 2007. Standard minimum quality standards for roasted and roasted and ground cofee - Characteristic: Gourmet Cofee. Federal Ofcial Gazette of Brazil.June,2007.Section1,24. [22] Brazil. Secretariat of Agriculture and Supply of Sao˜ Paulo State. Resolution No 19, Abril 5, 2010. Standard minimum quality standards for roasted and roasted and ground cofee - characteristic: Traditional cofee. Federal Ofcial Gazette of Brazil.Abril,2007.Section1,26. [23]P.D.Oliveira,F.M.Borem,E.P.Isquierdo,G.D.S.Giomo,´ R. R. de Lima, and R. A. Cardoso, “Physiological aspects of cofee beans, processed and dried through diferent methods, associated with sensory quality,” Cofee Science,vol.8,no.2,pp. 211–220, 2013. [24] A. J. Muller,¨ L. Huebner, and C. F. V. Souza, “Evaluation of physico-chemical quality of diferent brands of roasted cofee soluble and in powder marketed in the region of Vale do Taquari/RS,” Revista Brasileira de Tecnologia Agroindustrial,vol. 7, pp. 1004–1012, 2013. [25] M. A. Morgano, L. F. Pauluci, D. M. B. Mantovani, and E. E. M. Mory, “Mineral determination in green cofee,” Food Science and Technology,vol.22,pp.19–23,2002. [26] J. S. dos Santos, M. L. P. dos Santos, M. M. Conti, S. N. dos Santos, and E. de Oliveira, “Evaluation of some metals in Brazilian cofees cultivated during the process of conversion from conventional to organic agriculture,” Food Chemistry,vol. 115,no.4,pp.1405–1410,2009. [27] P. Liu, C.-N. Wang, X.-Y. Song, and Y.-N. Wu, “Dietary intake of lead and cadmium by children and adults - Result calculated from dietary recall and available lead/cadmium level in food in comparison to result from food duplicate diet method,” International Journal of Hygiene and Environmental Health,vol. 213, no. 6, pp. 450–457, 2010. [28] M. Grembecka, E. Malinowska, and P. Szefer, “Diferentiation of market cofee and its infusions in view of their mineral composition,” Science of the Total Environment,vol.383,no.1–3, pp.59–69,2007. [29] L. Noel,R.Chekri,S.Millouretal.,“Li,Cr,Mn,Co,Ni,Cu,Zn,¨ Se and Mo levels in foodstufs from the Second French TDS,” Food Chemistry,vol.132,no.3,pp.1502–1513,2012. [30] E. Stelmach, P. Pohl, and A. Szymczycha-Madeja, “Te suitabil- ity of the simplifed method of the analysis of cofee infusions on the content of Ca, Cu, Fe, Mg, Mn and Zn and the study of the efect of preparation conditions on the leachability of elements into the cofee brew,” Food Chemistry,vol.141,no.3,pp.1956– 1961, 2013. Hindawi Journal of Food Quality Volume 2018, Article ID 5967130, 8 pages https://doi.org/10.1155/2018/5967130

Research Article Effect of Yeast Fermentation of Green Coffee Beans on Antioxidant Activity and Consumer Acceptability

Han Sub Kwak ,1 Yoonhwa Jeong ,2 and Misook Kim 2

1 Division of Functional Food Research, Korea Food Research Institute, 245 Nongsaengmyeong-ro, Iseo-myeon, Wanju-gun, Jeollabuk-do 55365, Republic of Korea 2Department of Food Science and Nutrition, Dankook University, 19 Dandae-ro, 119 Dongnam-gu, Cheonan-si, Chungnam 31116, Republic of Korea

Correspondence should be addressed to Misook Kim; [email protected]

Received 16 November 2017; Accepted 23 January 2018; Published 6 March 2018

Academic Editor: Gabriel Henrique Horta de Oliveira

Copyright © 2018 Han Sub Kwak et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This study assessed the functionality and consumer acceptance of yeast fermented coffee beans. Green coffee beans were fermented for 24 h with three different yeast strains to increase functionality. The yeast fermentation was effective in fortifying the functionality of coffee by significantly increasing antioxidant activity according to the results of ORAC and SOD-like𝑃 assay( < 0.05). The TPC and TFC contents in the fermented coffee beans were significantly higher than those in the controls𝑃 ( < 0.05). The consumer acceptance for the fermented coffee beans was slightly lower than that of the controls. Fermentation seemed to influence the aroma and flavor of coffee. However, agglomerative hierarchical clustering analysis revealed that approximately 39% of consumers significantly liked one of the fermented coffees (F3) more than the controls (𝑃 < 0.05). These consumers indicated that the yeast fermentation of green coffee beans did not generate a negative aroma or flavor and can be attractive with high antioxidant activity.

1. Introduction The functionality of food products can be increased by fermentation. Ginseng is one of the best known functional Coffeeisoneofthemosttradedcommoditiesworldwide.It foods. Ginsenoside contents and SOD-like activity after is harvested in tropical regions and mostly exported to devel- fermentation were significantly increased by solid-microbial oped countries such as the USA, Europe, Russia, Japan, and fermentation of white ginseng [14, 15]. Increased antioxi- Korea [1, 2]. Many coffee drinkers consume coffee on a regular dant activities and phenolic compounds are also frequently basis, similar to tea, and coffee is regarded as a hedonic food. observed in fermented tea products. Jayabalan et al. [16] However, coffee contains significant amounts of phenolic showed an increase in epicatechin isomers over 18 days compounds such as chlorogenic and hydroxycinnamic acids of kombucha tea fermentation. DPPH radical scavenging and antioxidants including caffeine, melanoidins, and other activity and total polyphenol content in Pu-Erh tea were Maillard reaction products and volatile compounds [3–5]. significantly increased by fermentation [17]. Levels of chlorogenic and hydroxycinnamic acids are also The fermentation of coffee is known as coffee cherry determinedonthefinalaromaandtasteoftheroastedcoffee fermentation and effectively removes the mucilage layer prior [6, 7]. Richelle et al. [8] demonstrated that a cup of Robusta to the drying process for obtaining green coffee beans [18]. and Arabica coffee had two times more antioxidant activity Therefore, the primary objective of coffee cherry fermen- than a cup of green and black tea. A cup of coffee per day tation is to improve the ease of obtaining green coffee was found to reduce the relative risk of diabetes by 7% in beans rather than to increase the functionality of the coffee meta-analysis [9] and moderate coffee drinking (below 4 cups beans. Green coffee beans can gain higher functionality per day) showed the strongest inverse relation to heart failure with additional processing steps such as soaking in fruit [10]. In some European countries, coffee is one of the major extracts and fermentation. Lim et al. [19] reported higher sources of antioxidants in the human diet [11–13]. antioxidant activity, total polyphenol contents, and consumer 2 Journal of Food Quality acceptability after soaking green coffee beans in mulberry 2.3. Fermentation Characteristics. The pH of the fermenta- extract. However, the fermentation of green coffee beans tion solution was measured using an electronic pH meter has not been widely studied as a second processing step (Orion 3 star, Thermo Fisher Scientific Inc., Waltham, MA, for increasing antioxidant activity and phenolic compounds. USA) at 0 and 24 h. As the fermentation of tea products increases their antiox- The reducing sugar content in the green coffee beans idantactivityandthenumberofphenoliccompounds,the was determined by the DNS method [20] with some mod- antioxidant activity and phenolic compounds in green coffee ifications. A sample (200 uL) was mixed with 600 uL DNS ∘ beans could also be increased by fermentation. The fermented solutionandheatedat100C for 5 min. Water (3 mL) was coffee beans can become fortified functional coffee beans. added to the reaction solution and its absorbance was Therefore, the objective of this study was to ferment green measured at 550 nm. D-Glucose solutions (0.1–2.5 mg/mL) coffee beans using commercial yeasts and investigate the were used for a standard curve (𝑅 =0.997). resulting physicochemical properties, antioxidant activity, Yeast-containing solutions for green coffee bean fermen- total polyphenol and flavonoid contents, and consumer tation were serially diluted with 0.85% NaCl solution and acceptability with the intention of making fortified functional spread onto the surface of YPD agar plates. The spread ∘ coffee beans. plates were incubated for 24 h at 30 C to determine the total microbial counts. 2. Materials and Methods 2.4. Physicochemical Characteristics of Roasted Coffee. Mois- 2.1. Chemicals. Sodium carbonate (Na2CO3), sodium nitrite ture content determination was performed according to the (NaNO2), monosodium phosphate (NaH2PO4), disodium Association of Official Agricultural Chemists [21] guidelines. phosphate (Na2HPO4), potassium acetate (CH3COOK), and ∘ ⋅ Approximately 1 g of ground coffee beans was dried at 105 C aluminum chloride (AlCl3 6H2O) were purchased from Duk- after roasting until the weight became consistent. san Pure Chemicals (Ansan-si, Korea). HCl, NaCl, ethanol, and methanol were supplied by Samchun Pure Chemicals Crude ash content was also measured with the Associ- ation of Official Agricultural Chemists [21] guidelines. The (Pyeongtaek-si, Korea). D-Glucose was bought from Duchefa ∘ Biochemie BV (Haarlem, Netherlands). YPD was purchased roasted coffee beans (1 g) were incinerated at 550 Cuntilthe from Becton Dickinson (Sparks, MD, USA). Folin-Ciocal- weight became consistent. Crude ash content was determined 󸀠 teu’s phenol reagent, fluorescein sodium, 2,2 -azobis(2-am- gravimetrically. idinopropane)dihydrochloride (AAPH), 2,5-dinitrosalicylic The color of the ground coffee after roasting was mea- acid (DNS), trolox, ethylenediaminetetraacetic acid (EDTA), sured using a Hunter colorimeter system (JC-801S, Color pyrogallol, quercetin, and gallic acid were supplied by Sigma- Techno system Co., Tokyo, Japan). Ground coffee (2 g) was 𝐿 Aldrich LLC (St. Louis, MO, USA). put into a small Petri dish for measurement. Lightness ( ), redness (𝑎), and yellowness (𝑏)weremeasured.𝐿, 𝑎,and𝑏 valuesforthestandardwhitecolorplatewere𝐿 =98.38,𝑎 = 2.2. Coffee Fermentation, Roasting, and Brewing. Green coffee 𝑏 − beans (300 g; Brazil Ipanema Euro Natural, Coiners Inter- 0.29, and = 0.41, respectively. national Ltd., Bucheon-si, Korea) were put into a water 7 (450 mL) and yeast (3.75 mL, 1.0 × 10 CFU/mL) mixture. 2.5. Antioxidant Activity. Oxygen radical absorbance capac- Three different commercial yeastsSaccharomyces ( species) ity (ORAC) was measured using the method described by Ou were used for fermentation: F1 (Lalvin 71B, Lallemand Inc., et al. [22] with some modifications for in vitro antioxidant Montreal, Canada), F2 (Lalvin Cy3079, Lallemand Inc.), and activity. Phosphate (NaH2PO4-Na2HPO4) buffer (10 mM, pH F3 (BDX, Lallemand Inc.). Fermentation was conducted for 7.0) was used to dissolve fluorescein powder. Coffee extract ∘ 𝜇 𝜇 24 h at 30 C. After fermentation, green coffee beans were (50 L)and25mMfluoresceinsolution(150 L) were mixed washed three times using sterile water. The fermented coffee and then incubated in a dark room for 10 min. A volume ∘ 𝜇 beans were dried in a dry oven at 45 Cuntiltheirmoisture of 25 Lof120mMAAPHsolutionwasaddedtothecoffee content reached 10%. Two controls were used in this study; extract and fluorescein mixture. The control was 10 mM one control (C1) was the original green coffee beans and the phosphate buffer instead of coffee extract. Fluorescence was other was a control (C2) treated with the same procedure as measuredbythemicroplatereader(SpectramaxM2,Molec- theyeastfermentedcoffeebeansbutwithoutyeasts. ular Devices, LLC., Sunnyvale, CA, USA). Measurements Coffee roasting was conducted with an automatic coffee were taken every minute for 90 min (excitation wavelength: roastertoensureasimilarroastingconditionforallgreen 485 nm; emission wavelength: 535 nm). The ORAC values 𝜇 coffee beans (Behmor 1600, Behmor Inc., Incline Village, NV, were calculated by the following formula and presented as M 𝜇 USA). The green coffee beans were put into the roaster and trolox equivalent/mL of coffee ( M TE/mL): roasted at maximum heating power (1,600 W) for 16 min. ORAC (𝜇MTE/g) Theroastedcoffeebeanswerecooledfor15minandkeptat room temperature for three days. The roasted coffee beans 𝐶 ×(AUCSample − AUCBlank)×𝑘 (1) = Trolox , were ground using the medium option on a coffee grinder ( − ) (KG79, De’Longhi, Milano, Italy). The ground coffee (36 g) AUCTrolox AUCBlank 𝐶 𝑘 wasbrewedwith600mLoffilteredwaterusingacoffeemaker where Trolox , , and AUC were the concentration of trolox (HD7564, Philips, Netherlands). (5 𝜇M), the sample dilution factor, and the area under Journal of Food Quality 3

Table 1: Change of pH, reducing sugar, and viable cells before and after fermentation.

pH Reducing sugar (mg/mL) Viable cell count (log CFU/mL) Sample 0h 24h 0h 24h 0h 24h c(1) a a C2 5.98 5.13 0.68 ± 0.01 2.57 ± 0.13 <1.16 ± 0.53 7.78 ± 0.70 b a a F1 5.79 4.92 1.12 ± 0.02 2.51 ± 0.12 6.50 ± 5.83 9.31 ± 0.85 a a a F2 5.93 4.52 1.28 ± 0.03 2.47 ± 0.09 6.41 ± 5.62 9.25 ± 0.86 b a b F3 5.78 5.04 1.12 ± 0.01 2.67 ± 0.09 6.30 ± 5.30 9.03 ± 0.81 (1) Different superscripts within a column meant significant difference at 𝑃<0.05 by Fisher’s least significant difference test. the curve, respectively. AUC was calculated based on the at 510 nm against the blank sample using a microplate reader following formula: (Spectra max M2, Molecular Devices, LLC). The difference in absorbance between the sample and the blank was compared 90 𝑓𝑛 totheabsorbanceofthequercetinsolutionusedasthe AUC =1+∑ , (2) 𝑅2 𝑛=1 𝑓0 positive control ( = 0.999). The amount of flavonoids in the coffee was presented as mg quercetin equivalent/mL of coffee where 𝑓𝑛 is the fluorescence at time 𝑛 (min). extract. Superoxide dismutase-like (SOD-like) activity was determined using the method described by S. Marklund and 2.7. Consumer Acceptance Test. The consumer acceptance G. Marklund [23] with some modifications. Coffee extract test for the yeast fermented coffee was performed with 74 (400 𝜇L), Tris-HCl buffer (600 𝜇L, 50 mM tris(hydroxym- subjects (24 males and 50 females, ages 19–30) who were ethyl)aminomethane and 10 mM ethylenediaminetetraacetic students and staff at Dankook University (Yongin-si, Korea). acid (EDTA), pH 8.0), and 7.2mM pyrogallol (40 𝜇L) were Subjects were coffee drinkers that consumed Americano ∘ mixed together and kept at 25 C for 10 min. The reaction was coffee (espresso coffee extract + hot water) at least once a stopped by adding 0.1 N HCl (20 𝜇L).Theabsorbancewas week. Consumer testing was conducted in an open area under measured at 420 nm using a UV/visible spectrophotometer incandescent lighting. Subjects were assigned to tables and (Ultrospec 2100 Pro, Biochrom Ltd., Cambridge, UK). SOD- were prohibited from talking and using cellphones during the like activity was calculated based on the following equation: test.Coffee(20mL)wasservedin60mLpapercupsusinga completedbalanceddesign[26]inordertominimizecarry- 𝐴 over effects. The temperature of the coffee was approximately SOD-like activity (%) = (1− ) × 100, (3) ∘ 𝐵 60 C. Acceptance of overall quality, coffee aroma, bitter taste, astringent taste, and smooth mouthfeel was rated using a 𝐴 𝐵 where is absorbance of the sample and is absorbance of nine-point hedonic scale on paper ballots. Each category the control. on the scale was labeled with numbers and descriptors in order to give a clear understanding of the scale. Starting from 2.6. Total Polyphenol and Flavonoid Contents. The total pol- the left side, the scale was labeled as 1 = dislike extremely, yphenol content (TPC) of brewed coffee was measured using 2 = dislike very much, 3 = dislike moderately, 4 = dislike Singleton’s method [24] with some modifications. Coffee slightly,5=neitherlikenordislike,6=likeslightly,7=like extract (20 𝜇L) was diluted with 1,580 𝜇Lofdistilledwater. moderately, 8 = like very much, and 9 = extremely like. A cup Diluted coffee (160 𝜇L) was mixed with Folin-Ciocalteu’s of filtered water was served as a palate cleanser. Subjects were phenol reagent (10 𝜇L) and allowed to sit for 8 min. A volume monetarily compensated after the test (approximately $8.5). of 30 𝜇Lof20%Na2CO3 solution was added and the mixture was incubated in a dark room for 2 h. Distilled water instead 2.8. Statistical Analyses. Results were analyzed using Excel of coffee extract was used as a control. The absorbance was integrated statistical software (XLSTAT version 2012, Addin- measured at 765 nm by a microplate reader (Spectra max M2, soft, Paris, France) in this study. Analysis of variance Molecular Devices, LLC.). Gallic acid solutions (0-1 mg/mL) (ANOVA) was performed to identify significant differences 2 were used to generate a standard curve (𝑅 =0.997).The across samples. Post hoc analysis was carried out using results were presented as mg gallic acid equivalent/mL of Fisher’s least significant test at 𝑃 < 0.05 when the significance coffee extract (mg GAE/mL). wasobservedbyANOVA.Agglomerativehierarchicalclus- The total flavonoid content (TFC) for each coffee extract tering (AHC) analysis was conducted to generate consumer was evaluated according to the method described by Dewanto segment groups based on their overall quality ratings. study with some modifications [25]. Coffee extract (250 𝜇L), 𝜇 distilled water (1 mL), and 75 L of 5% NaNO2 were mixed 3. Results and Discussion together. After 5 min, 10% AlCl3⋅6H2Osolution(150𝜇L) was added and incubated for 6 min. 1 N NaOH (500 𝜇L) was 3.1. Fermentation Characteristics. Measurements of pH val- injected, and the mixture was incubated for 11 min. A blank ues, reducing sugar contents, and viable cell counts were sample was substituted for the diluted coffee extract with conducted to investigate the effect of yeast fermentation on distilled water. The absorbance of the sample was measured green coffee beans (Table 1). Reducing sugar contents for 4 Journal of Food Quality

Table 2: Moisture content, crude ash, and color of ground roasted coffee beans after 24 h of yeast fermentation.

Color Sample Moisture content (%) Crude ash (%) 𝐿𝑎𝑏Δ𝐸 a(1) a a a a b C1 1.50 ± 0.09 4.73 ± 0.05 31.52 ± 1.01 7. 9 1 ± 0.44 16.78 ± 1.92 69.60 ± 1.27 b cd c b c a C2 1.30 ± 0.05 4.08 ± 0.06 25.24 ± 0.62 4.90 ± 0.25 8.70 ± 0.89 74.00 ± 0.62 b d ab a ab b F1 1.27 ± 0.17 4.05 ± 0.06 30.62 ± 0.81 7. 1 1 ± 0.94 14.74 ± 0.80 69.91 ± 0.52 b c c b c a F2 1.23 ± 0.06 4.15 ± 0.05 25.17 ± 0.68 4.43 ± 0.26 7. 6 3 ± 0.53 73.90 ± 0.61 b b b a b b F3 1.31 ± 0.10 4.40 ± 0.05 29.78 ± 0.97 6.97 ± 0.54 13.75 ± 1.08 70.51 ± 0.70 (1) Different superscripts within a column meant significant difference at P < 0.05 by Fisher’s least significant difference test.

50 100 79.79ab 34.09a 82.76a 34.95a 37.44a 69.58b 40 80

20.25b 30 60 14.11b M/mL of coffee) M/mL of

 33.67c 29.63c 20 40 SOD-like activity (%) activity SOD-like 10 20 ORAC (TE ORAC

0 0 C1 C2 F1 F2 F3 C1 C2 F1 F2 F3 (a) (b)

Figure 1: Antioxidant activities of yeast fermented coffee extracts by oxygen radical absorbance capacity (a) and superoxide dismutase-like (SOD-like) activity (b) assays. Different letters meant significant difference at P < 0.05 by Fisher’s least significant difference test.

C2, F1, F2, and F3 were 2.567, 2.507, 2.469, and 2.673 mg. than those of C2 and the fermented coffee beans𝑃 ( < 0.05). pH decreased after 24 h of fermentation. The pH for C2 This was due to slight overdrying prior to roasting. The ash also decreased. This decrease would be due to the soluble content was also significantly higher in C1 at 4.73%, while organic acids released from green coffee beans [27] and theothersrangedfrom4.05to4.40%.Moisturecontentand produced by the fermentation of naturally occurring lactic crude ash showed a strong negative correlation (𝑅 =0.894, acid bacteria in coffee beans. After 24 h of fermentation, there 𝑃 = 0.041). The color of the ground roasted coffee beans was no significant difference in reducing sugar content in the showed differences across the samples. C1, F1, and F3 had fermented solution (𝑃 > 0.05). Viable cell counts for C2, higher 𝐿, 𝑎,and𝑏 values and lower Δ𝐸 values than C2 and F1, F2, and F3 showed that fermentation was conducted by F2. The color of the ground coffee beans showed a different yeasts (Table 1). The initial number of viable microbial cells pattern in comparison with the moisture content and crude was less than 1.16 log CFU/mL. The number of microbial ash. Therefore, the different colors of the roasted coffee beans cells, naturally present in green coffee beans, increased to likely partially originated from the fermentation as well as the 7.78 CFU/mL after 24 h fermentation. Solutions from F1, F2, difference in moisture contents. and F3, fermented by starter culture, showed approximately 2.5 higher log CFU/mL of solution than that from C2. 3.3. Antioxidant Activity. The antioxidant activities of the This provided supporting evidence for the proliferation of coffee extracts in ORAC and SOD-like assays are shown yeasts during fermentation and the role of yeasts in the in Figure 1. ORAC for the yeast fermented coffee extracts fermentation. Throughout the 24 h of fermentation, there was were 34.95, 37.44, and 34.09 𝜇M TE/mL for F1, F2, and F3, no drastic change in pH and reducing sugar content; hence respectively (Figure 1(a)). ORAC for C1 and C2 were 20.25 yeast fermentation occurred across all samples. and 14.11 𝜇M TE/mL, respectively. There was no significant differenceinORACvaluesaftersoakingthegreencoffee 3.2. Quality Characteristics of Fermented Coffee Beans. The beans for 24 h (𝑃 > 0.05); however, a slight decrease in moisture content, crude ash, and color of fermented coffee the ORAC value was observed in C2. It seemed that the beans after roasting are presented in Table 2. No signifi- soluble antioxidants in green coffee beans might be eluted cant difference across fermented coffee beans and C2 was into water. The ORAC values of the yeast fermented coffee observed. This indicates that the drying and roasting of the extracts (F1, F2, and F3) were significantly higher than those fermented green coffee beans were performed identically. The from the controls (C1 and C2) (𝑃 < 0.05). There was moisture content of C1 was 1.50% and was significantly higher no significant difference across the fermented coffee beans Journal of Food Quality 5

1.5 1.5 1.30a 1.21a 1.11b 1.18ab 1.01bc 1.04b 1.0 0.86c 1.0 0.89cd 0.72c 0.77d

0.5 0.5 TFC (mM coffee) of QE/mL TPC (GAE coffee) of mg/mL

0.0 0.0 C1 C2 F1 F2 F3 C1 C2 F1 F2 F3 (a) (b) Figure 2: Total polyphenol (a) and flavonoid (b) contents of yeast fermented coffee extracts. Different letters meant significant difference at P < 0.05 by Fisher’s least significant difference test.

(𝑃 < 0.05). In SOD-like activity, F1, F2, and F3 showed In TFC, the yeast fermented coffee extracts (F1, F2, 82.76, 79.79, and 69.58%, respectively (Figure 1(b)). On the and F3) had 1.01, 1.04, and 1.21 QE mg/mL of coffee extract, other hand, C1 and C2 showed 33.67 and 29.63% SOD-like respectively, and F3 had significantly higher𝑃 ( < 0.05) activity, respectively, and were not significantly different𝑃> ( TFC than F1 and F2 (Figure 2(b)). C1 and C2 had lower 0.05). An increase of more than two times the SOD-like amounts of TFC than the yeast fermented coffee extracts. activity was observed after 24 h of yeast fermentation. Among Yeast fermentation was effective in increasing the number of the fermented samples, F1 showed the highest activity and flavonoids in the coffee extracts. The increase in TFC might the activity was significantly higher than F3𝑃 ( < 0.05). be due to the conversion of insoluble phenolic compounds Soaking green coffee beans in water (C2) did not significantly into soluble flavonoids during fermentation [31]. The ratio of influence the antioxidant activity of the coffee extracts𝑃> ( TFC to TPC in C1 and C2 was 1.03 and 1.07, while F1, F2, 0.05) although the ORAC value of C2 was slightly lower and F3 had ratios of 0.91, 0.88, and 0.93, respectively. The than that of C1. Fermentation for 24 h was effective in ratio was lower in the yeast fermented coffee extracts. Fer- increasing the antioxidant activity significantly. The increase mentation was more effective in producing soluble phenolic of antioxidant activity by fermentation was similar to other compounds than flavonoids. This result contrasted with those fermented materials such as ginseng, garlic, tea, and soy for fermented Columbian coffee [32]. However, Columbian [15, 16, 28, 29]. coffee fermentation was conducted with broken green beans, which might elute more soluble phenolic compounds easily 3.4. Total Polyphenol and Flavonoid Contents. The TPC and over seven days of fermentation. This would also influence TFC of the yeast fermented coffee extracts are presented in the roasting process. Figure 2. The yeast fermented coffee extracts (F1, F2, and F3) had 1.11, 1.18, and 1.30 GAE mg/mL of coffee extract, 3.5. Consumer Acceptance Test. Consumer acceptance for the respectively (Figure 2(a)). F3 had significantly higher TPC fermented and control coffee extracts is shown in Table 3. than F1 (𝑃 < 0.05). Control samples (C1 and C2) had Mean overall quality rating was the highest in C1, followed significantly lower amounts of TPC than the yeast fermented by F3. F1 and F2 had significantly lower overall quality coffee extracts (𝑃 < 0.05). C2 had the lowest TPC at ratings than C1 at 4.35 and 4.43, respectively (𝑃 < 0.05). 0.72 GAE mg/mL. This was due to the elution of soluble C2 also received slightly lower overall quality ratings than phenolic compounds in the green coffee beans during the C1. Therefore, the soaking process does appear to slightly soaking period (24 h). A similar finding was observed after influence the coffee quality. Color acceptance was not sig- soaking green coffee beans in mulberry extract [19]. In Lim’s nificantly different (𝑃 < 0.05), which meant that there study, the TPC of the coffee extracts decreased after 6 h of was no difference in the appearance of the coffee extract. soaking. Despite the decrease in TPC due to the soaking Aroma acceptance ratings for the fermented coffee samples process in this study, fermentation positively influenced TPC (F1, F2, and F3) were significantly lower than those for in the coffee extracts. During the fermentation process, the controls (C1 and C2) (𝑃 < 0.05). The consumer strongly bound phenolic compounds in the cell wall may acceptability was within the range of the previous acceptance become weakened [30], making them easier to extract after study using various roasting conditions [33]. Fermentation roasting. Phenolic compounds in coffee are mostly in the might negatively influence the aroma acceptance of coffee. forms of chlorogenic acids with 5-O-caffeoyl-quinic acid such Across the fermented coffee samples (F1, F2, and F3) the as caffeic, ferulic, p-coumaric, and caffeoylquinic acid [3]. differences in consumer acceptability might be related to the Feruloylquinic acid, di-caffeoyl-quinic acid, and proantho- presence of different volatile compounds in coffee beans after cyanidins are also detected in coffee [3]. fermentation. Different yeasts showed different flavor profiles 6 Journal of Food Quality

Table 3: Overall quality and acceptances of color, aroma, sourness, bitterness, astringency, and mouthfeel of yeast fermented coffee extract by 74 consumers.

Sample Overall quality Color Aroma Sourness Bitterness Astringency Mouthfeel a(1) a a a a a C1 5.09 ± 1.72 5.59 ± 1.76 5.62 ± 1.63 5.12 ± 1.86 4.51 ± 1.90 4.81 ± 1.80a 5.80 ± 1.75 ab a a a ab ab ab C2 4.69 ± 1.65 5.30 ± 1.75 5.12 ± 1.70 5.07 ± 1.49 4.12 ± 1.94 4.57 ± 1.82 5.35 ± 1.68 b a b a b b b F1 4.35 ± 1.70 5.18 ± 1.48 4.08 ± 1.85 4.78 ± 1.50 3.84 ± 1.76 4.20 ± 1.76 4.99 ± 1.72 b a b b ab b b F2 4.43 ± 1.46 5.23 ± 1.44 4.51 ± 1.51 4.19 ± 1.58 4.01 ± 1.85 4.03 ± 1.79 4.85 ± 1.66 ab a b a ab ab b F3 4.84 ± 1.67 5.54 ± 1.60 4.45 ± 1.95 4.84 ± 1.47 4.41 ± 1.89 4.36 ± 1.58 5.12 ± 1.57 (1) Different superscripts within a column meant significant difference at P < 0.05 by Fisher’s least significant difference test.

Table 4: Agglomerative hierarchical clustering (AHC) analysis of overall quality.

(1) Class Subject C1 C2 F1 F2 F3 b(2) b b a a 1283.79 ± 1.20 B(3) 3.82 ± 1.61 B 3.57 ± 1.48 B 4.89 ± 1.66 A 5.32 ± 1.61 A a b bc d cd 2435.95± 1.53 A 5.23 ± 1.49 A 4.81 ± 1.72 A 4.09 ± 1.29 B 4.51 ± 1.71 B (1) (2) Three subjects were removed from AHC analysis because they marked same ratings for entire samples. Different superscripts within a row meant (3) significant difference at 𝑃<0.05 by Fisher’s least significant difference test. Different subscripts within a column meant significant difference at 𝑃<0.05. in rice distilled liquor [34] and cachac¸a [35]. Acceptance a significant increase in antioxidant activity, TPC, and TFC. ratings for sourness, bitterness, astringency, and mouthfeel Yeast fermentation of green coffee beans causes bound phe- were lower in the yeast fermented coffee samples than in C1 nolic compounds to be released after roasting. The consumer and C2. acceptance for the fermented coffee beans was slightly lower In order to segment consumers into a small number of than for the controls. Fermentation might negatively influ- groups, AHC analysis was performed using the overall quality ence the aroma and flavor of coffee extracts. However, the ratings for all except three consumers that selected the same consumer segmentation revealed that approximately 39.4% of acceptance ratings for the entire sample (Table 4). Consumers consumers preferred one of the fermented coffees (F3) more were divided into two clusters generated by AHC analysis. than the controls. Therefore, it can be concluded that yeast Cluster 1 was composed of 39.4% of the total consumers. fermentation did not always generate a negative aroma and These were the consumers who preferred fermented coffee flavor for consumers. If fermentation was carried out with (F2 and F3), while disliking the controls (C1 and C2) (𝑃< properly selected yeasts, fermented coffee can be attractive 0.05). On the other hand, cluster 2 was composed of 60.6% to coffee consumers, and coffee manufacturers can diversify of consumers. These consumers had higher overall quality theirproductswithhigherfunctionality. ratings for the controls (C1 and C2). They also rated F1 as an average of 4.81 and had the least acceptance for F2 at 4.09. Conflicts of Interest Approximately 40% of consumers preferred the fermented coffee extracts (F2 and F3) although fermented coffee had The authors declare that they have no conflicts of interest. lower mean overall quality ratings for the entire group of consumers. Therefore, the fermented coffee samples (F2 and F3) did not have lower acceptance ratings for all consumers; References they were acceptable and showed higher acceptability for [1] H. D. Vieira, “Coffee: The plant and its cultivation,” in Plant- 39.4% of consumers. This result provides supporting evidence Parasitic Nematodes of Coffee, M. Souza, Ed., pp. 3–18, Springer, that fermentation does not have a negative influence on Dordrecht, Netherlands, 2008. the consumer acceptance of coffee. Consumers are therefore [2] “International Coffee Organization (ICO),” 2015, http://www segmented and another coffee product can be created for .ico.org/trade statistics.asp. approximately 40% of coffee consumers. In addition, the high [3] P.Parras, M. Mart´ınez-Tome,´ A. M. Jimenez,´ and M. A. Murcia, antioxidant activity and phenolic compounds in fermented “Antioxidant capacity of coffees of several origins brewed coffee can be attractive to those consumers with moderate following three different procedures,” Food Chemistry,vol.102, acceptance for fermented coffee. The consumer acceptance no.3,pp.582–592,2007. would also increase with awareness of the high antioxidant [4] M. Perez-Mart´ ´ınez, B. Caemmerer, M. P. de Pena,˜ C. Con- activity and phenolic compounds content as shown in a cepcion,´ and L. W. Kroh, “Influence of brewing method and previous blind and informed consumer acceptance test for acidity regulators on the antioxidant capacity of coffee brews,” blueberry functional beverages [36]. Journal of Agricultural and Food Chemistry,vol.58,no.5,pp. 2958–2965, 2010. 4. Conclusion [5] I. Sanchez-Gonz´ alez,´ A. Jimenez-Escrig,´ and F. Saura-Calixto, “In vitro antioxidant activity of coffees brewed using different Yeast fermentation of green coffee beans for 24 h was effec- procedures (Italian, espresso and filter),” Food Chemistry,vol. tive in fortifying the functionality of coffee by inducing 90,no.1-2,pp.133–139,2005. Journal of Food Quality 7

[6]J.R.Oliveira-Neto,S.G.Rezende,C.DeFatima´ Reis, S. R. [21] AOAC., Official Methods of Analysis, Association of Official Benjamin, M. L. Rocha, and E. De Souza Gil, “Electrochemical Analytical Chemists, Gaithersburg, MD, USA, 18th edition behavior and determination of major phenolic antioxidants in edition, 2005. selected coffee samples,” Food Chemistry,vol.190,no.1,Article [22] B. Ou, M. Hampsch-Woodill, and R. L. Prior, “Development and ID 17651, pp. 506–512, 2016. validation of an improved oxygen radical absorbance capacity [7]J.A.Vignoli,M.C.Viegas,D.G.Bassoli,andM.D.T.Benassi, assay using fluorescein as the fluorescent probe,” Journal of “Roasting process affects differently the bioactive compounds Agricultural and Food Chemistry,vol.49,no.10,pp.4619–4626, and the antioxidant activity of arabica and robusta coffees,” Food 2001. Research International,vol.61,no.1,pp.279–285,2014. [23] S. Marklund and G. Marklund, “Involvement of the superoxide [8] M. Richelle, I. Tavazzi, and E. Offord, “Comparison of the anion radical in the autoxidation of pyrogallol and a convenient antioxidant activity of commonly consumed polyphenolic bev- assay for superoxide dismutase,” European Journal of Biochem- erages (coffee, cocoa, and tea) prepared per cup serving,” istry,vol.47,no.3,pp.469–474,1974. JournalofAgriculturalandFoodChemistry,vol.49,no.7,pp. [24] V. L. Singleton, R. Orthofer, and R. M. Lamuela-Raventos,´ 3438–3442, 2001. “Analysis of total phenols and other oxidation substrates and [9] R. Huxley, C. M. Y. Lee, F. Barzi et al., “Coffee, decaffeinated antioxidants by means of folin-ciocalteu reagent,” Methods in coffee, and tea consumption in relation to incident type2 Enzymology, vol. 299, pp. 152–178, 1999. diabetes mellitus: a systematic review with meta-analysis,” [25] V. Dewanto, X. Wu, and R. H. Liu, “Processed sweet corn has rchives of Internal Medicine,vol.169,no.22,pp.2053–2063, higher antioxidant activity,” JournalofAgriculturalandFood 2009. Chemistry,vol.50,no.17,pp.4959–4964,2002. [10]E.Mostofsky,M.S.Rice,E.B.Levitan,andM.A.Mittleman, [26]H.J.MacFie,N.Bratchell,K.Greenhoff,andL.V.Vallis, “Habitual coffee consumption and risk of heart failure a dose- “Designs to balance the effect of order of presentation and first- response meta-analysis,” Circulation: Heart Failure,vol.5,no.4, ordercarry-overeffectsinhalltests,”Journal of Sensory Studies, pp.401–405,2012. vol. 4, no. 2, pp. 129–148, 1989. [27] C. I. Rodrigues, L. Marta, R. Maia, M. Miranda, M. Ribeir- [11] R. Pulido, M. Hernnndez-Garc˜ ´ıa, and F.Saura-Calixto, “Contri- inho, and C. Maguas,´ “Application of solid-phase extraction bution of beverages to the intake of lipophilic and hydrophilic to brewed coffee caffeine and organic acid determination by antioxidants in the Spanish diet,” European Journal of Clinical UV/HPLC,” Journal of Food Composition and Analysis,vol.20, Nutrition,vol.57,no.10,pp.1275–1282,2003. no. 5, pp. 440–448, 2007. [12] S. A. Qureshi, A. C. Lund, M. B. Veierød et al., “Food items [28] Y.-H. Pyo, T.-C. Lee, and Y.-C. Lee, “Effect of lactic acid contributing most to variation in antioxidant intake; a cross- fermentation on enrichment of antioxidant properties and sectional study among Norwegian women,” BMC Public Health, bioactive isoflavones in soybean,” Journal of Food Science,vol. vol. 14, no. 1, article no. 45, 2014. 70, no. 3, pp. S215–S220, 2005. [13]A.Svilaas,A.K.Sakhi,L.F.Andersenetal.,“Intakesof [29] E. Sato, M. Kohno, H. Hamano, and Y. Niwano, “Increased antioxidants in coffee, wine, and vegetables are correlated with anti-oxidative potency of garlic by spontaneous short-term plasma carotenoids in humans,” Journal of Nutrition,vol.134, fermentation,” Plant Foods for Human Nutrition,vol.61,no.4, no. 3, pp. 562–567, 2004. pp.157–160,2006. [14] B.-G. Kim, S.-Y. Choi, M.-R. Kim, H. J. Suh, and H. J. [30] M. Shi, Y. Yang , Q. Wang , Y. Zhang , Y. Wang , and Z. Zhang , Park, “Changes of ginsenosides in Korean red ginseng (Panax “Production of total polyphenol from fermented soybean curd ginseng) fermented by Lactobacillus plantarum M1,” Process residue by Lentinus edodes,” International Journal of Food Biochemistry,vol.45,no.8,pp.1319–1324,2010. Science & Technology,vol.47,no.6,pp.1215–1221,2012. [15] S.-I. Lim, C.-W. Cho, U.-K. Choi, and Y.-C. Kim, “Antioxidant [31] Y.-K. Lee, S.-I. Lee, J.-S. Kim et al., “Antioxidant activity of activity and ginsenoside pattern of fermented white ginseng,” greenteafermentedwithMonascuspilosus,”Journal of Applied Journal of Ginseng Research, vol. 34, no. 3, pp. 168–174, 2010. Biological Chemistry,vol.55,no.1,pp.19–25,2012. [16] R. Jayabalan, S. Marimuthu, and K. Swaminathan, “Changes in [32] J.-Y. Shin, H. Kim, D.-G. Kim, G.-H. Baek, H.-S. Jeong, and content of organic acids and tea polyphenols during kombucha K.-W. Yu, “Pharmacological activities of coffee roasted from tea fermentation,” Food Chemistry,vol.102,no.1,pp.392–398, fermented green coffee beans with fungal mycelia in solid- 2007. state culture,” Journal of the Korean Society of Food Science and Nutrition,vol.42,no.3,pp.487–496,2013. [17] K.-C. Jeng, C.-S. Chen, Y.-P.Fang, R. C.-W.Hou, and Y.-S.Chen, “Effect of microbial fermentation on content of statin, GABA, [33] L. C. Mendes, H. C. De Menezes, M. Aparecida, and A. P. and polyphenols in Pu-erh tea,” Journal of Agricultural and Food Da Silva, “Optimization of the roasting of robusta coffee (C. Chemistry,vol.55,no.21,pp.8787–8792,2007. canephora conillon) using acceptability tests and RSM,” Food Quality and Preference, vol. 12, no. 2, pp. 153–162, 2001. [18] C. F.Silva, L. R. Batista, L. M. Abreu, E. S. Dias, and R. F.Schwan, [34]H.S.Kwak,J.S.Seo,Y.Huretal.,“Influenceofyeast “Succession of bacterial and fungal communities during natural strains on the physicochemical characteristics, methanol and coffee (Coffea arabica) fermentation,” Food Microbiology,vol. acetaldehyde profiles and volatile compounds for Korean rice 25,no.8,pp.951–957,2008. distilled spirit,” JournaloftheInstituteofBrewing,vol.121,no.4, [19] H. H. Lim, S. Ji, H. S. Kwak et al., “Quality characteristics of pp. 574–580, 2015. coffee brewed from green beans soaked in mulberry (Morus [35] E. S. Oliveira, H. M. A. B. Cardello, E. M. Jeronimo, E. L. Rocha bombycis) extract,” JournaloftheKoreanSocietyofFoodScience Souza, and G. E. Serra, “The influence of different yeasts on and Nutrition,vol.44,no.4,pp.579–585,2015. the fermentation, composition and sensory quality of cachac¸a,” [20] T. K. Ghose, “Measurement of cellulase activities,” Pure and World Journal of Microbiology and Biotechnology,vol.21,no.5, Applied Chemistry,vol.59,no.2,pp.257–268,1987. pp. 707–715, 2005. 8 Journal of Food Quality

[36] M. K. Kim and H. S. Kwak, “Influence of functional information on consumer liking and consumer perception related to health claims for blueberry functional beverages,” International Jour- nal of Food Science & Technology,vol.50,no.1,pp.70–76,2015. Hindawi Journal of Food Quality Volume 2018, Article ID 7909467, 8 pages https://doi.org/10.1155/2018/7909467

Research Article Shade Trees Spatial Distribution and Its Effect on Grains and Beverage Quality of Shaded Coffee Trees

Francisco José da Silva Neto ,1 Kátia Priscilla Gomes Morinigo,1 Nathalia de França Guimarães ,2 Anderson de Souza Gallo ,2 Maicon Douglas Bispo de Souza,1 Rubismar Stolf,1 and Anastácia Fontanetti 1

1 Department of Rural Development, Universidade Federal de Sao˜ Carlos (UFSCar), Rodovia Anhanguera, Km 174, Araras, SP, Brazil 2Department of Soil, Institute of Agronomy, Universidade Federal Rural do Rio de Janeiro (UFRRJ), BR-465, Km 7, Seropedica,´ RJ, Brazil

Correspondence should be addressed to Francisco Jose´ da Silva Neto; [email protected] and Anastacia´ Fontanetti; [email protected] Received 30 September 2017; Accepted 27 December 2017; Published 25 February 2018

AcademicEditor:FernandoM.Botelho

Copyright © 2018 Francisco Jose´ da Silva Neto et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Shading coffee trees has gained importance, especially among smallholders, as an option to improve the products’ quality, therefore acquiring place at the specialty coffee market, where consumers are willing to give bonus for quality. This work aims to evaluate the influence of shade trees’ spatial distribution among coffee trees’ agronomic characteristics, yield, and beans and cup quality of shaded coffee trees. The experimental design consisted of completely randomized blocks with six repetitions and four treatments: coffee trees on shade trees planting rows, distant one meter from the trunk; coffee trees on shade trees planting row, distant six meters from the trunk; and coffee plants between the rows of shade trees, parallel to the previous treatments. The parameters analyzed were plant height, canopy diameter, plagiotropic branches’ length, yield, coffee fruits’ phenological stage, ripe cherries’ Brix degree, percentage of black, unripe, and insect damaged beans, bean size, and beverage quality. Shade trees quickened coffee fruits’ phenological stage of coffee trees nearest to them. This point also showed the best beverage quality, except for overripe fruits. The remaining parameters evaluated were not affected by shade trees’ spatial distribution.

1. Introduction biennial bearing [6]. It also protects coffee plants from strong winds, rains, or hail [7] and increases nutrient cycling and Brazilian production processes of specialty coffee have grown soil organic matter [8]. This can significantly increase crop in the past decades due to the increase of demand from inter- production stability. national markets [1]. The concept of specialty coffee is broad Shade trees also impact quality because maturation is and involves characteristics such as superior beverage qual- more uniform under shade [9]. Coffee beans’ size is enlarged ity, rare varieties, and location of cultivation but also can be as a consequence of fewer flowers and, also, the existence related to ecological, economic, or social sustainability [2]. of less fruits per plant once coffee trees are shaded [6]. This Still, coffee quality is the result of complex interactions be- benefic effect is possible in sites under optimal [10] and sub- tween the environment, management, and plant genetics [3]. optimal [11] conditions for coffee plants. However, Bossel- The implantation of shade trees in coffee plantation can mann et al. [12] found a decrease in coffee quality when shade bring about many benefits to the agroecosystem, such as trees were present in plantations at high altitudes. temperature reduction of air, soil, and leaf surface [4], as However, conditions in shaded coffee plantations are not well as the thermal amplitude [5], and softening the effect of spatially stable, especially in systems with lower rates of soil 2 Journal of Food Quality

D6L D1L

1 m (0.8 m) 6 m

3.5 m (5) (3) (1) D6E D1E

14 m

Shade trees planting row D1L: coffee trees distant 1 m from the trunk (6) (4) (2) N D6L: coffee trees distant 6 m from the trunk Shade trees between rows (1) Anadenanthera falcata; (2) Albizia polycephala; D1E: coffee trees distant 1 m from the trunk (3) Cassia grandis; (4) Cassia grandis; D6E: coffee trees distant 6 m from the trunk (5) Cassia grandis; (6) Anadenanthera falcata (a) (b) Figure 1: Graphical representation of the shaded coffee plantation (a); detailing of the treatments (b). Santo Antonioˆ do Jardim, SP, 2015.

Table 1: Climate data observed during the months in which the experiment was conducted. Santo Antonioˆ do Jardim, SP, 2015.

∘ RD PREC (mm) T ( C) RH (%) Months Total Total Average Min Max Average Min Max Average Feb 14 348 12.4 19.9 28.3 24.1 80 94 87 Mar 18 475 15.3 18.1 26.6 22.3 87 98 92 Apr 4 165 5.5 17.5 27.2 22.3 70 93 81 May 7 123 4.0 14.6 23.1 18.8 88 98 93 Jun 3 27 0.9 14.7 24.4 19.5 89 95 92 Jul 2 30 1.0 14.5 24.4 19.4 90 95 92 Aug 1 10 0.3 15.0 25.7 20.3 78 92 85 RD: rain days; PREC: precipitation; T: temperature; RH: relative air humidity. Source: meteorological station of the Retiro Santo Antonioˆ farm (2015). cover by trees. In the farthest points from the shade trees, The species used were Anadenanthera falcata, Albizia poly- microclimate conditions are rather similar to the ones in full- cephala, and Cassia grandis with the arrangement of 15.0 × −1 sun systems [13]. 14.0 meters (44 plants ha ) (Figure 1). Therefore, this work aims to evaluate the influence of Fertilization was carried out according to soil chemical shade trees’ spatial distribution on growth and productivity characteristics and Raij et al. [16] recommendation. Protected of coffee trees as well as on coffee quality. urea was applied three times a year, in the dosage of 50 g per plant, the last applications being done in November and 2. Materials and Methods December 2014 and February 2015; potassium chloride was applied twice a year, 10 g per plant, the last one being realized 2.1. Research Site. The research took place between April in February 2015; chicken litter and straw of coffee were andAugust2015atRetiroSantoAntonioˆ farm, located in applied once a year in the dosage of 1 kg per plant, the last ∘ 󸀠 Santo Antonioˆ do Jardim, Sao˜ Paulo state, Brazil (22 06 Sand applied in October 2014. No fertilization with phosphorus or ∘ 󸀠 46 40 W,850 m above sea level). The climatic conditions were boron was used during the experiment. categorized by Koppen¨ [14] as Cwb, temperate with dry win- ters and warm summers. Maximum and minimum tempera- 2.2. Treatments and Experimental Design. The experiment ture and air relative humidity are shown in Table 1, as well as design was of completely randomized blocks with six repe- rainfall volumes. Soil was classified as Red-Yellow Argisol titions and four treatments, spatially distributed due to shade [15]. Soil chemical characteristics are described in Table 2. trees’ localization: coffee trees on shade trees planting rows, A site implanted in 2007 with the coffee variety of Obata˜ distant one meter from the trunk (D1L); coffee trees on shade Vermelho and the arrangement of 0.8 × 3.5 meters (3,570 trees planting rows, distant six meters from the trunk (D6L); −1 plants ha ) was used. Shade trees were implanted in 2009. coffee plants between the rows of shade trees, parallel to Journal of Food Quality 3

Table 2: Soil chemical characterization in depths of 0,0–0,20 m and 0,20–0,40 cm of shaded coffee plantations. Santo Antonioˆ do Jardim, SP, 2015. Depth Unit 0,0–0,20 m 0,20–0,40 m pH CaCl2 5.12 5.20 3 PMg⋅dm 36.13 30.00 3 M⋅Og⋅dm 23.96 21.92 −3 Kmmolc⋅dm 2.96 3.06 −3 Ca mmolc⋅dm 29.80 24.94 −3 Mg mmolc⋅dm 13.16 11.28 −3 H+Al mmolc⋅dm 27.54 34.63 −3 Al mmolc⋅dm 1.54 2.48 −3 SB mmolc⋅dm 45.94 39.29 −3 CEC mmolc⋅dm 73.48 73.89 V % 63.60 54.88 SB:sumofbases;CEC:cation-exchangecapacity;V:basesaturation. treatments D1L and D6L (D1E and D6E) (Figure 1). The plot Posteriorly, cherries were allocated separately in plastic consisted of three consecutive coffee trees. net bags and naturally dried in the sun over a slab of concrete until 13% of humidity was reached. Then, coffee beans were 2.3. Evaluations mechanically peeled (Carmomaq, Tecnologia e Inovac¸ao˜ para aIndustria´ do Cafe,´ model: DRC 1, number: 9498, year: 2.3.1. Phenological Evolution. To determine the ideal time to 2012); in this operation, the parchment was also removed and harvest, the phenological evolution of coffee cherries was once more weighed. Immediately after, beans’ humidity was tracked once the endosperm expansion has finished. Pez- determined with a benchtop grain moisture tester (Moisture zopane et al.’s [17] methodology was adapted to contemplate and Purity Tester G650, Gehaka Agri). only cherry phenological evolution, (1) green, (2) yellowish green, (3) cherry, and (4) overripe. Four productive branches, 2.3.4. Coffee Quality. Samplesof300gofgreencoffeebeans two in each exposition side to the sun, were marked with a of each plot and each cherry stage (unripe, ripe, and overripe) plasticribbonsothesamecherrieswerealwaysevaluated.The were analyzed by the Qualicafex Specialty Coffees company. There, the percentage of defect (black, unripe, and insect evaluations took place on April 24th, May 1st and 14th, and damaged beans), sieve retention, and cup quality were July 28th. assessed. Cup quality test follows the SCCA (Specialty Coffee Association of America) guidelines and the notes are given 2.3.2. Coffee Trees’ Agronomical Characteristics and Brix De- according to Brazilian law (Brazil, 2003). In Brazil, Arabic gree. Coffee trees’ agronomical characteristics were evalu- coffee quality is divided into seven subgroups. In order from ated previous to harvesting. Plant height, given in meters, the best for wort quality, the subgroups are (i) strictly soft: was measured from the insertion of the orthotropic branch coffee that has beverage with all the characteristics of aroma in the ground to the apical bud. Canopy diameter, in meters, and taste of the soft beverage, but strongly accentuated; (ii) was measured perpendicular to planting rows, measuring soft: coffee beverage with smooth, sweet, and pleasant taste the farthest distance between the first pair of leaves from and aroma; (iii) softish: coffee beverage with weakly sweet opposites plagiotropic branches. Finally, plagiotropic branch and smooth, but still with no signs of astringency; (iv) hard: length per plant, given in centimeters, was the mean of ten coffee beverage with acrid and astringency taste, but still random branches. withnosignsofoddflavors;(v)Rioysh:coffeebeverage Brix degree evaluation used the juice of three ripe cherries with soft flavor, slightly resembling iodoform; vi) Rio: coffee per plot. Cherries were squished in the optical reader of a beverage with typical flavor of iodoform; (vii) Rio Zona: ∘ portable refractometer. The results were given in Bx; each coffee beverage with strongly accentuated aroma and taste, degree corresponds to one gram of sucrose in 100 grams of strongly resembling iodoform or phenic acid and repugnant solution. flavor [18]. To determine the percentage of defect, a subsample of 2.3.3. Coffee Yield. Harvest was conducted manually by strip- 100 g had the defects picked up manually and weighed. The picking between July 28 and 30, when more than half of percentages were given dividing the mass of each defect by the the cherries in the hole experiment were ripe. Cherries were initial mass. Sieve retention also used 100 g subsample and, washed, to remove impurities, and separated into unripe similarly,thepercentagewasgivenbydividingtheretained fruits (UF), ripe fruits (RF), and overripe fruits (OF). Each mass by the original mass. Samples were passed through group of cherries was weighed, in a semianalytic balance sieves 13 (5.15 mm mesh) and 17 (6.75 mm mesh). machine, separately and together, so the total weight was revealed. Cherries volume was measured with a 2,000 ml 2.4. Statistical Analyses. The data expressed in percentage graduated cylinder. was transformed using the function 𝑦=arcsin(√𝑥/100).The 4 Journal of Food Quality percentage of defects, sieve retention, plant height, canopy 4 D1L diameter, plagiotropic branch length, and yield were submit- 3 ted to ANOVA test, and when they were significantly distinct the means were compared with the Tukey test at 𝑃 ≤ 0.05. 2 The remaining data was analyzed descriptively. 1 3. Results and Discussion 4 D6L 3 Coffee trees’ plant height, canopy diameter, and plagiotropic branch length did not differ between the treatments (Table 3). 2 Such a fact, probably, is due to low variation in light intensity 1 among the systems. Ricci et al. [4] related that in systems 4 with low soil cover and little variation in shadows’ levels, D1E such as coffee trees shaded with Erythrina,whichcoveronly 3 2 to 6% of the soil, external morphological changes are not scores Phenological observedincoffeetrees.Pezzopaneetal.[19]evaluatingcoffee 2 trees shaded with banana trees also did not find differences in 1 plants’ heights on points among the systems, closer or farther 4 from shade trees. D6E However,thesameauthorsfounddifferencesincanopy 3 diameters of coffee trees closer to the banana trees. Ricci et al. [20] observed increases in plant height and canopy diameter 2 of coffee trees in shaded systems when compared with full- 1 sun ones. The greater plant height and canopy diameter rep- 18/4 25/4 02/5 09/5 16/05 23/5 30/05 resent the coffee tree’s effort to compensate for the less light Figure 2: Coffee fruits’ phenological stages evolution of shaded availability under shade, as an attempt to reach solar energy coffee plants due to shade trees’ spatial distribution. Santoonio Antˆ [21]. do Jardim, SP, 2015. Phenological scores: (1) green, (2) yellowish The treatment closest to shade trees (D1L) promoted green, (3) cherry, and (4) overripe. D1L: coffee trees on shade trees faster maturation of cherries and the treatment farther from planting rows, distant one meter from the trunk; D6L: coffee trees shade trees (D6E) was the latest to ripen (Figure 2). This on shade trees planting rows, distant six meters from the trunk; D1E: result differs from the one mentioned by Ricci et al. [22] coffee plants between the rows of shade trees, parallel to treatment that reported later maturation in Obata˜ coffee in agroforest D1L; D6E: coffee plants between the rows of shade trees, parallel to systems with Erythrina and banana trees. But this result is treatment D6L. similar to the one found by Lunz et al. [9] that showed later and nonuniform maturation when solar exposition of coffee cherries of Obata˜ was greater due to distancing of coffee trees Treatments also did not affect Brix degree of ripe cherries. In the closest treatment to shade trees, D1L, the mean was from shade trees. ∘ ∘ 19.8 Bx, the intermediate points D6L and D1E showed 21.1 Bx Regardless of that, maturation speed is reduced by shad- ∘ and 19.2 Bx, respectively, and in D6L, the furthest point from ow [10], being an important factor for cup quality increase. ∘ This is a result, mostly, of reduction of average temperature, shade trees was 18.3 Bx. Similar values were found by Silva as is the case in high “mountain” coffee, which presents a et al. [28]. However, those authors consider Brix degree as a better cup quality [10]. So, it is possible that maturation ends poor tool to predict cup quality. early not because it was sped up, but due to early flowering. The total yields of cherries and of unripe, ripe, and According to DaMatta et al. [23], the flower bud remains overripe cherries in mass and volume were not affected by dormant until the, so-called, “blossom showers” when flowers shade trees’ spatial distribution (Table 4). There was not any finish their growth and blossom. Therefore, in a point close to difference in the treatment of green coffee beans in total or in shade trees, soil moisture may have been conserved as shown any of the maturation stages of cherries (Table 5). by Dhanya et al. [24], and the accumulation of smaller rainfall Those results contradict the ones found by Raij et al. may have been enough to stimulate blossom at these points. [16] that identified loss of yield in coffee trees near shading banana trees. Yield in this work was greater than the Brazilian ThisiscorroboratedbyLunzetal.[9]thatidentifiedlatter −1 blossoming in more distant coffee plants of shade trees. average, of 22.49 sacks ha [29].Itisworthwhiletohighlight that, in highly technified systems of shaded coffee, researches Another hypothesis for the early maturation could be −1 competition for water among coffee trees and shade trees due in Brazil showed high yields, such as 141 sacks ha in to shared soil volume by roots of both species in D1L, such as coffee plants shaded by Swietenia macrophylla in the Federal thecaseobservedbyCoelhoetal.[25].AccordingtoMorais District. On the other hand, in native forest coffee yields were −1 et al. [26], hydric stress associated with high temperature alittleover4sacksha in Minas Gerais [30]. can quicken cherries maturation. However, Morinigo [27], No differences among the treatments were found regard- working in the same site, period, and plots, did not find any ing the percentages of defects (Table 6). Unripe harvested differenceinsoilmoisturein0,0–20and0,20–0,40mdepths. cherries presented a greater percentage of unripe bean defect, Journal of Food Quality 5

Table 3: Plant height (m), canopy diameter (m), and plagiotropic branch length (cm) of coffee trees due to shade trees spatial distribution. Santo Antonioˆ do Jardim, SP, 2015.

Plant height Canopy diameter Branch length Treatments mcm ns ns ns D1L 1.49 1.45 19.56 D6L 1.46 1.43 17.12 D1E 1.70 1.50 19.36 D6E 1.57 1.41 18.11 VC% 14.85 9.51 8.58 ns: nonsignificant according to 𝐹 test at 5% of significance. D1L: coffee trees on shade trees planting rows, distant one meter from the trunk; D6L: coffeetrees on shade trees planting rows, distant six meters from the trunk; D1E: coffeelants p between the rows of shade trees, parallel to treatment D1L; D6E: coffeeplants between the rows of shade trees, parallel to treatment D6L. VC (%): variation coefficient.

Table 4: Total coffee fruits production after harvesting and production of unripe fruits (UF), ripe fruits (RF), and overripe fruits (OF)after washing and fruit separation due to shade trees spatial distribution. Santo Antonioˆ do Jardim, SP, 2015.

UF RF OF Total (UF + RF + OF) Coffee fruits −1 kg⋅plant ns ns ns ns D1L 0.489 1.262 1.009 2.760 D6L 0.676 1.267 1.072 3.015 D1E 0.744 1.260 0.680 2.684 D6E 0.845 1.680 0.866 3.391 VC% 46.07 37.04 38.79 34.18 Volume per plant −1 L⋅plant ns ns ns ns D1L 0,850 1,983 1,900 4,733 D6L 1,178 2,044 1,967 5,189 D1E 1,244 2,039 1,183 4,467 D6E 1,400 2,811 1,356 5,567 VC% 41,80 35,95 47,96 35,42 ns: nonsignificant according to 𝐹 test at 5% of significance. D1L: coffee trees on shade trees planting rows, distant one meter from the trunk; D6L: coffeetrees on shade trees planting rows, distant six meters from the trunk; D1E: coffeelants p between the rows of shade trees, parallel to treatment D1L; D6E: coffeeplants between the rows of shade trees, parallel to treatment D6L. VC (%): variation coefficient. which was expected. However, that defect was recurrent in Muschler [11] observed that, in pollarded systems with all cherries stages of maturation, which is the same as the Erythrina as a shade tree, coffee beans were greater than in results found by Carvalho et al. [31]. Black beans were also systems with heavier shadows. The author then hypothesizes found in all treatments; still they were more significant in that closer to the shade tree coffee beans were bigger enough unripe harvested cherries. The fermentation that causes these to bring the mean up and benefit the system as a whole. How- defects is greater in unripe cherries, once they have higher ever, the same effect was not observed in this work. water content [32]. Unripe harvested cherries showed a greater percent of All treatments presented low percentages, close to zero, small beans (Table 7). This can be explained by the immatu- of insect damaged beans, 0.3% for D1L, 0.0% for D6L, 1.0% rity of cherries. The endosperm is filled until approximately for D1E, and 0.3 for D6E (Table 6). According to Beer et al. the 17th week after blossoming [23], and during the cherries [5], trees may be the habitat to natural enemies that control separation by maturation, all immature fruits are considered the borer. Besides, the same authors report experiences where unripe, even the ones that have not finished endosperm fill- the entomopathogenic fungi Beauveria bassiana,whichwere ing. inoculated in the research site, had better persistence due to Besides, defective beans are smaller than those of high shade trees. quality [34]. So, beans coming from unripe cherries, the ones The variable bean size, represented by the sieve retention, that showed the most defects, should be smaller. alsohadnoeffectonthetreatmentsduetoshadetrees’spatial To ripe cherries, the worst cup quality was found in D6E, distribution (Table 7). There is consensus in the literature as hard, while the remaining treatments presented beverage thatcoffeebeansenlargeduetoshade[9–12,22,33].Still, softish to D6L and soft to D1L and D1E (Table 8). This result 6 Journal of Food Quality

Table 5: Green coffee production due to shade trees spatial distribution. Santoonio Antˆ do Jardim, SP, 2015.

UF RF OF Total (UF + RF + OF) −1 kg⋅plant ns ns ns ns D1L 0.084 0.232 0.358 0.674 D6L 0.122 0.217 0.355 0.694 D1E 0.125 0.228 0.236 0.589 D6E 0.150 0.288 0.253 0.691 VC% 43.61 44.67 37.97 33.53 −1 sacks⋅ha ns ns ns ns D1L 5.00 13.79 21.30 40.10 D6L 7.28 12.91 21.10 41.30 D1E 7.42 13.54 14.02 35.00 D6E 8.94 17.16 14.45 40.60 VC% 43.68 44.68 37.89 33.50 −1 kg⋅ha ns ns ns ns D1L 299.96 827.28 1277.82 2405.07 D6L 436.85 774.74 1265.92 2477.68 D1E 445.18 812.40 840.97 2098.56 D6E 536.25 1029.64 867.16 2433.04 VC% 43.68 44.68 37.89 33.50 ns: nonsignificant according to 𝐹 test at 5% of significance. D1L: coffee trees on shade trees planting rows, distant one meter from the trunk; D6L: coffeetrees on shade trees planting rows, distant six meters from the trunk; D1E: coffeelants p between the rows of shade trees, parallel to treatment D1L; D6E: coffeeplants between the rows of shade trees, parallel to treatment D6L. VC (%): variation coefficient.

Table 6: Percentage of black, unripe, and insect damaged beans in samples of unripe fruits (UF), ripe fruits (RF), and overripe fruits (OF) of shaded coffee trees due to shade trees spatial distribution. Santo Antonioˆ do Jardim, SP, 2015.

Black beans Unripe beans Insect damaged beans UF RF OF UF RF OF UF RF OF % ns ns ns ns ns ns ns ns ns D1L 14 0 0 29 4 4 0 0 1 D6L 17 1 0 22 5 3 0 0 0 D1E 15 1 1 35 5 5 0 0 3 D6E 14 0 1 26 5 4 1 0 0 VC (%) 14.9 23.7 25.3 7.8 30.0 33.4 13.9 0.0 34.3 ns: nonsignificant according to 𝐹 test at 5% of significance. D1L: coffee trees on shade trees planting rows, distant one meter from the trunk; D6L: coffeetrees on shade trees planting rows, distant six meters from the trunk; D1E: coffeelants p between the rows of shade trees, parallel to treatment D1L; D6E: coffeeplants between the rows of shade trees, parallel to treatment D6L. VC (%): variation coefficient.

Table 7: Percentage of coffee beans of size smaller than sieve 13 and greater than sieve 17 of shaded coffee trees’ unripe, ripe, and overripe fruits due to shade trees spatial distributions. Santo Antonioˆ do Jardim, SP, 2015.

Beans < 13 Beans > 17 UF RF OF UF RF OF % ns ns ns ns ns ns D1L 45 11 14 15 19 25 D6L 52 16 12 11 18 25 D1E 53 16 17 15 9 21 D6E 42 14 13 11 9 21 VC (%) 3.8 26.4 15.0 33.4 30.9 19.6 ns: nonsignificant according to 𝐹 test at 5% of significance. D1L: coffee trees on shade trees planting rows, distant one meter from the trunk; D6L: coffeetrees on shade trees planting rows, distant six meters from the trunk; D1E: coffeelants p between the rows of shade trees, parallel to treatment D1L; D6E: coffeeplants between the rows of shade trees, parallel to treatment D6L. VC (%): variation coefficient. Journal of Food Quality 7

Table 8: Beverage scores of samples from unripe fruits (UF), ripe fruits (RF), and overripe (OR) fruits of shaded coffee trees due to shade trees spatial distribution. Santo Antonioˆ do Jardim, SP, 2015.

Cup quality UF RF OF D1L Just soft Soft Hard D6L Hard Just soft Soft D1E Hard Soft Soft D6E Hard Hard Soft D1L: coffee trees on shade trees planting rows, distant one meter from the trunk; D6L: coffee trees on shade trees planting rows, distant six meters fromthe trunk; D1E: coffee plants between the rows of shade trees, parallel to treatment D1L; D6E: coffee plants between the rows of shade trees, parallel to treatment D6L. was expected once cup quality is well correlated with shad- References ows [10]. Also, in a more distant point from shade trees, microclimate may be similar to full-sun systems [13]. [1] C. G. Z. Teixeira and T. S. Milagres, “Economicidade e certif- Among overripe cherries, the worst quality was found in icac¸ao˜ da cafeicultura familiar na Zona da Mata de Minas Ger- ais,” Pesquisa Agropecuaria´ Tropical,vol.39,no.4,pp.317–329, D1L, as hard, and the remaining treatments presented bever- 2009. ageassoft.Onceagain,thiswasexpected,sincematuration was over first in these treatments (Figure 2). Those cherries [2]S.P.Pereira,G.FBartholo,andP.T.G.Guimaraes,˜ “Cafes´ espe- ciais: iniciativas brasileiras e tendenciasˆ de consumo,” Serie´ were, probably, exposed to more adverse weather, since they Documentos,vol.41,2004. stayed longer after maturation. Increasing the exposure to ¨ ¨ opportunist microbes and possibility of fermentation while [3] P. Laderach, T. Oberthur, S. Cook et al., “Systematic agronomic farm management for improved coffee quality,” Field Crops the cherries are still attached to the coffee tree affects quality Research,vol.120,no.3,pp.321–329,2011. [35]. Iamanaka et al. [36] found that cherries that are overripe [4] M. S. F. Ricci, D. G. Cocheto Junior, and F. F. D. Almeida, “Con- and still attached to the coffee plant presented the worst dic¸oes˜ microclimaticas,´ fenologia e morfologia externa de ca- beverage quality and greater infestation of pathogenic fungi. feeiros em sistemas arborizados e a pleno sol,” Coffee Science, But when the cherries were overripe and do not stay long vol. 8, no. 3, pp. 379–388, 2013. exposed to weather, they did not differ from those harvested [5] J.Beer,R.Muschler,D.Kass,andE.Somarriba,“Shademanage- on plane maturation [37]. This explains the best quality of ment in coffee and cacao plantations,” Agroforestry Systems,vol. overripe cherries on D6L and D1E. According to Silva et al. 38,no.1-3,pp.139–164,1998. [28], the harvested unripe cherries present the worst cup [6] F.M.DaMatta,“Ecophysiologicalconstraintsontheproduction quality, as found in this work, as they are just as soft to D1L ofshadedandunshadedcoffee:Areview,”Field Crops Research, andhardtotheremainingtreatments. vol. 86, no. 2-3, pp. 99–114, 2004. [7] A. P. Alvarenga, R. S. Vale, L. Couto, F. A. F. Vale, and A. B. 4. Conclusion Vale, “Aspectos fisiologicos´ da cultura do cafeeseupotencial´ produtivo em sistemas agroflorestais,” Agrossilvicultura,vol.1, Shade trees’ spatial distribution affects the phenological evo- no. 2, pp. 195–202, 2004. lution of cherries and cup quality of coffee trees. Maturation [8] M.M.Campanha,R.H.Santos,G.B.Freitas,H.E.Martinez,C. was accelerated in coffee plants closer to shade trees, probably Jaramillo-Botero, and S. L. Garcia, “Analise´ comparativa das due to anticipated blossom. Werecommend monitoring since caracter´ısticasdaserrapilheiraedosoloemcafezais(Coffeaara- the blossoming phase to elucidate the results presented here. bica L.) cultivados em sistema agroflorestal e em monocultura, Also, the best cup quality was obtained in coffee beans coming na Zona da Mata MG,” Revista Arvore´ ,vol.31,no.5,pp.805–812, from coffee trees closer to shade trees, except for overripe 2007. cherries. [9] A.M.P.Lunz,M.S.Bernardes,C.A.Righi,J.D.Costa,J.L.Fava- rin, and J. G. Cortez, “Qualidade do cafear´ abica´ em sistema agroflorestal de seringueira (Hevea brasiliensis Muell.¨ Arg.) e Conflicts of Interest em monocultivo in,” in IV Simposio´ de Pesquisa dos Cafes´ do Brasil, Anais...Londrina, Embrapa Cafe,´ 2005. The authors declare that they have no conflicts of interest [10]P.Vaast,B.Bertrand,J.-J.Perriot,B.Guyot,andM.Genard,´ regarding the publication of this paper. “Fruit thinning and shade improve bean characteristics and beverage quality of coffee (Coffea arabica L.) under optimal Acknowledgments conditions,” Journal of the Science of Food and Agriculture,vol. 86,no.2,pp.197–204,2006. The authors profusely thank Mr. Jefferson Adorno, the owner [11] R. G. Muschler, “Shade improves coffee quality in a sub-optimal of Retiro Santo Antonioˆ farm, his family, and staff for the coffee-zone of Costa Rica,” Agroforestry Systems,vol.51,no.2, opportunity to learn with their innovative experience at pp.131–139,2001. Kayna˜ Coffee and for all the essential collaborations to the ex- [12] A. S. Bosselmann, K. Dons, T. Oberthur, C. S. Olsen, A. Raebild, ecution of this work. and H. Usma, “The influence of shade trees on coffee quality 8 Journal of Food Quality

in small holder coffee agroforestry systems in Southern Colom- [29] Quarto Levantamento, vol. 2015, 2016, http://www.conab.gov bia,” Agriculture, Ecosystems & Environment,vol.129,no.1-3,pp. .br/OlalaCMS/uploads/arquivos/15 12 17 09 02 47 boletim 253–260, 2009. cafe d440. [13]J.R.M.Pezzopane,M.M.S.Marsetti,W.R.Ferrari,andJ.E.M. [30]C.Jaramillo-Botero,H.E.P.Martinez,andR.H.S.SANTOS, Pezzopane, “Microclimatic alterations in a conilon coffee crop “Caracter´ısticas do cafe´ (Coffea arabica L.) sombreado no norte grown shaded by green dwarf coconut trees,” Revista Cienciaˆ da America´ Latina e no Brasil: analise´ comparativa,” Coffee Agronomicaˆ , vol. 42, no. 4, pp. 865–871, 2011. Science,vol.1,no.2,pp.94–102,2006. [ 14 ] W. K a n d W. K oppen,¨ Climatologia: con un estudio de los climas [31]A.Carvalho,R.S.Garrutti,A.A.Teixeira,L.M.Pupo,andL.C. de tierra, Mexico: Fondo de cultura economica´ , Fondo de cultura Monaco, “Ocorrenciaˆ dos principais defeitos do cafeemv´ arias´ economica,´ Mexico, 1 edition, 1948. fases de maturaçao˜ dos frutos,” Bragantia, vol. 29, no. unico, pp. [15]N.d.Guimaraes,A.d.Gallo,A.Fontanettietal.,“Biomassae˜ 207–220, 1970. atividade microbiana do solo em diferentes sistemas de cultivo [32] A. A. Custodio,´ N. M. Gomes, and L. A. Lima, “Efeito da irri- do cafeeiro,” Revista de Cienciasˆ Agrarias´ ,vol.40,no.1,pp.34– gaçao˜ sobre a classificaçao˜ do cafe,”´ Engenharia Agr´ıcola,vol.27, 44, 2017. no.3,pp.391–701,2007. [16]B.vanRaij,H.Cantarella,J.A.Quaggio,andA.M.C.Furlani, [33] A. J. J. Souza, S. N. Matsumoto, M. R. Malta, and R. J. Guimaraes,˜ “Recomendaçao˜ de Adubaçao˜ e Calagem para o Estado de Sao˜ “Qualidade do cafe´ arborizado e a pleno sol, em manejo pos-´ Paulo, Instituto Agronomicoˆ - Fundaçao˜ IAC,” in Recomen- colheita no sudoeste da Bahia,” Coffee Science,vol.8,no.2,pp. daçao˜ de AdubaçaoeCalagemparaoEstadodeS˜ ao˜ Paulo,p. 109–120, 2013. 285, Instituto Agronomicoˆ - Fundaçao˜ IAC, Campinas, 1996. [34] J. C. F. Mendonça,A.S.Franca,andL.S.Oliveira,“Physical [17]J.R.Pezzopane,M.J.PedroJunior,´ R. A. Thomaziello, and M. characterization of non-defective and defective Arabica and B.Camargo,“Escalaparaavaliaçao˜ de estadios´ fenologicos´ do Robusta coffees before and after roasting,” Journal of Food cafeeiro arabica,”´ Bragantia,vol.62,no.3,pp.499–505,2003. Engineering,vol.92,no.4,pp.474–479,2009. [18] Brazil. Ministerio´ da Agricultura, Pecuaria´ e Abastecimento. [35] G. S. Giomo, “Uma boa pos-colheita´ e´ segredo da qualidade,” A Intruçao˜ Normativa n. 8, de 11 de junho de 2003. Diario´ oficial Lavoura,vol.115,no.688,pp.12–21,2012. da Uniao,˜ poder Executivo, Bras´ılia, 13 jun. 2003. [36] B. T. Iamanaka, A. A. Teixeira, A. R. R. Teixeira, M. V. Copetti, [19] J. R. Pezzopane, M. J. Pedro Junior,´ P.B. Gallo, M. B. Carmargo, N. Bragagnolo, and M. H. Taniwaki, “Reprint of “The mycobiota and L. C. Fazuoli, “Avaliaçoes˜ fenologicas´ e agronomicasˆ em of coffee beans and its influence on the coffee beverage”,” Food cafear´ abica´ cultivado a pleno sol e consorciado com banana Research International,vol.61,pp.33–38,2014. ’Prata Ana’,”˜ Bragantia,vol.66,no.4,pp.701–709,2007. [37] R. d. Garruti and A. G. Gomes, “Influenciaˆ do estado de matu- [20] E. Vivier, D. H. Raulet, A. Moretta et al., “Innate or adaptive raçao˜ sobreˆ a qualidade da bebida do cafenaregi´ ao˜ do vale do immunity? The example of natural killer cells,” Science,vol.331, para´ıba,” Bragantia, vol. 20, no. unico, pp. 989–995, 1961. no.6013,pp.44–49,2011. [21] J.I.Fahl,M.L.C.Carelli,J.Vega,andA.C.Magalhaes,“Nitrogen and irradiance levels affecting net photosynthesis and growth of young coffee plants (Coffea arabica L.),” Journal of Horticultural Science,vol.69,no.1,pp.161–169,1994. [22]M.d.Ricci,J.R.Costa,A.N.Pinto,andV.L.Santos,“Cultivo organicodecultivaresdecafˆ e´ a pleno sol e sombreado,” Pesquisa Agropecuaria´ Brasileira,vol.41,no.4,pp.569–575,2006. [23] F.M.DaMatta,C.P.Ronchi,M.Maestri,andR.S.Barros,“Eco- physiology of coffee growth and production,” Brazilian Journal of Plant Physiology,vol.19,no.4,pp.485–510,2007. [24]B.Dhanya,B.N.Sathish,S.Viswanath,andS.Purushothaman, “Ecosystem services of native trees: Experiences from two traditional agroforestry systems in Karnataka, Southern India,” International Journal of Biodiversity Science, Ecosystem Services & Management, vol. 10, no. 2, pp. 101–111, 2014. [25] R. A. Coelho, S. N. Matsumoto, C. L. Lemos, and F. A. Souza, “N´ıvel de sombreamento, umidade do solo e morfologia do cafeeiro em sistemas agroflorestais,” Revista Ceres,vol.57,no. 1, pp. 95–102, 2010. [26] H. Morais, P. H. Caramori, M. S. Koguishi, J. C. Gomes, and A. M. Ribeiro, “Sombreamento de cafeeiros durante o desenvolvimento das gemas florais e seus efeitos sobre a frutificaçao˜ e produçao,”˜ Cienciaˆ Rural,vol.39,no.2,pp.400–406,2009. [27] K. P. Morinigo, N. D. Guimaraes,˜ R. Stolf et al., “Efeitos da dis- tribuiçao˜ de arvores´ sobre atributos do solo em cafeeiro sombreado,” Coffee Science,vol.12,no.4,p.517,2017. [28] S. D. A. Silva, D. M. de Queiroz, F. D. A. C. Pinto, and N. T. San- tos, “Coffee quality and its relationship with Brix degree and col- orimetric information of coffee cherries,” Precision Agriculture, vol. 15, no. 5, pp. 543–554, 2014. Hindawi Journal of Food Quality Volume 2018, Article ID 3285452, 7 pages https://doi.org/10.1155/2018/3285452

Research Article Propositions on the Optimal Number of Q-Graders and R-Graders

Lucas Louzada Pereira ,1 Rogério Carvalho Guarçoni,2 Gustavo Soares de Souza,2 Dério Brioschi Junior,3 Taís Rizzo Moreira ,4 and Carla Schwengber ten Caten1

1 Departamento de Engenharia de Produc¸aoeTransportes,UniversidadeFederaldoRioGrandedoSul(UFRGS),˜ ∘ AvenidaOsvaldoAranha99,5 Andar, 90035-190 Porto Alegre, RS, Brazil 2Capixaba Institute for Technical Assistance, Research and Rural Extension, Department of Statistics, Rua Afonso Sarlo 160, Bento Ferreira, 29052-010 Vitoria,´ ES, Brazil 3Federal Institute of Esp´ırito Santo, Food Science and Technology Program, Rua Elizabeth Minete Perim, S/N, Bairro Sao˜ Rafael, 29375-000 Venda Nova do Imigrante, ES, Brazil 4Center of Agricultural Sciences and Engineering, Federal University of Esp´ıritoSanto,AltoUniversitario,´ S/N, Guararema, 29500-000 Alegre, ES, Brazil

Correspondence should be addressed to Lucas Louzada Pereira; [email protected]

Received 14 September 2017; Accepted 13 November 2017; Published 12 February 2018

Academic Editor: Gabriel Henrique Horta de Oliveira

Copyright © 2018 Lucas Louzada Pereira et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Sensory analysis or cup testing has been widely used in the coffee production chain for the validation of final quality. The tasters are responsible for defining the patterns and qualitative profiles of the drink based on the sensorial analysis and according totheir gustatory sensibilities, which are often acquired by professionalperience. ex However, the literature has not discussed in detail the relationship between the number of tasters and the consistency of sensorial analysis. Thus, using the bootstrap simulation methodology to estimate the optimum plot size, this study quantifies and proposes a specific number of tasters for the process of sensorial analysis of specialty coffees. The results indicate that the use of 6 tasters is sufficient to conduct sensorial analysis following SCA and BSCA protocol for coffees in the Arabica group, as well as 6 tasters for coil and Conilon coffees. From this number, no gains in precision are observed in the process of sensorial analysis of coffee with addition tasters.

1. Introduction The taster tends to prefer one sensory profile tothe detriment of others or even according to commercial and Inthecoffeeindustry,thetastingprocedureisusedto industrial standards in order to meet the demand of certain negotiatethepricebasedonthequalityofthedrink,which clients. This sensorial classification of coffees is based on is described by the tasters using personal opinion and tasting thetasteofthecup.Tastingtechniquemayvaryand/orbe experience accumulated over the years [1]. Although the modifiedbasedonthelocationofthetasting[4].Thefield tasting process is widely used, for Di Donfrancesco et al. [2], of sensory evaluation of coffees grew rapidly in the second this is not the best method to evaluate coffee quality, due to a half of the twentieth century, along with the expansion of the range of factors that interfere with the tasting process. processed food and consumer products industries. Sensory AccordingtoAlvaradoandLinnemann[3],the“taster” evaluation is now an irreplaceable tool in the food industry is a judge who performs the sensory evaluation, and this and is used in key sectors in the production of specialty agentisinchargeofevaluatingthequalityofthecoffeeand, coffees [5]. consequently, the evaluation of quality is influenced by his However, there is no consensus in the literature about the sensorial perceptions. number of testers to be used during sensory analysis of coffee, 2 Journal of Food Quality as well as the terminology for the profession. Traditionally, of Analysis and Research in Coffee, LAPC, of the Federal judges should be screened for sensory capabilities that meet Institute of Esp´ırito Santo in 2016. All the coffees (Arabica and the requirements of assessments [6]. In companies that have Conilon) come from the 2015/2016 harvest. tasters, it is common that the team consists of a quality The roasting was carried out with a roaster Laboratto control leader and assistants; however, in many producing TGP2, and all the samples were toasted between 8 and 10 regions (origins/rural areas) the purchasing offices use only minutes. For the standardization of the roasting process, the one person in this role. set of Agtron-SCA disks, the degree of roast of these samples, It is known that tasters trained and experienced in was chosen among the colors determined by the disks #65 sensorial analysis of coffee are not common because the and #55, for both Arabica coffee and Conilon coffee. The method is not usually taught in colleges and technical schools roasting process was performed 24 hours in advance, and in Brazil. For Dzung and Dzuan [5], one of the main problems after the roasting and cooling the samples remained sealed. in using experts in sensory evaluation is that the qualification The grinding was performed after the time of 24 hours of of the tasters is not well defined. In accordance with ISO 856- rest after the roasting process according to the methodology 2, experience is not the only criterion of a specialist; he must of sensorial analysis established by the SCA. The grinding also be trained and have high sensory sensitivity. was carried out in electric mill model BUNN G3, with Some methodologies, such as the Specialty Coffee Associ- granulometry in 20 meshes, following the US standard. ation (SCA), Brazil Specialty Coffee Association (BSCA) pro- For the evaluation of the Arabica coffees, the protocols tocols for Arabica coffee (Coffea arabica L.), and the Uganda of the SCA [11] and BSCA [12] were used. For the samples Coffee Development Authority (UCDA) for Conilon coffee of Conilon coffee, the same procedure of roasting, resting, (Coffea canephora Pierre ex Froehner), define procedures and and grinding of Arabica coffee was established, but the UCDA provide protocols for sensory evaluation of specialty coffees. protocol was used for sensory evaluation. These methodologies are commonly adopted in Brazil and the rest of the world for coffee quality contests and scientific 2.2. Method of Sampling. The quality of the Arabica coffee studies with applications of sensorial analysis. was evaluated using the SCA protocol, and it is expressed The use of trained judges is an important part of the through a centesimal numerical scale. The tasting form pro- sensory evaluation tradition [6], but for Ross [7] the cost vides an opportunity to evaluate eleven important attributes of sensory tests is very expensive, when a large number of for coffee: fragrance/aroma, uniformity, clean cup, sweetness, specialists is used due to the difficulty of accessing these flavor, acidity, body, aftertaste, balance, and overall and total professionals. note. Highly positive results arise from the perception of a The amount of tasters used in sensorial analysis of coffee balanced set formed by the evaluated attributes. The attributes may compromise the quality of the study. On the one hand, of Arabica coffee are the same in the two protocols, SCA and the use of few testers can cause loss of accuracy of the analysis. BSCA. On the other hand, the use of many tasters can be expensive The results of this sensory evaluation are established from becauseitusesagreaternumberoftastersthannecessary. a scale of 16 (sixteen) units representing quality levels with In addition, an environment with many people can cause intervals of 0.25 (one-quarter of a point) between numerical external noise, thus compromising the quality of the study, values between “6” and “10.” Coffees were considered good as already demonstrated by Pereira et al. [8]. from 6.00 to 6.75, very good from 7.00 to 7.75, excellent Therefore, the definition of the number of trained coffee from 8.00 to 8.75, and exceptional from 9.00 to 10.00 points. tasters has not been discussed in studies to define the The same interval procedure applies to all three protocols. optimal amount of coffee used in sensory analysis. Aiming Theoretically, a scale varies from a minimum value of 0to for greater consistency in the implementation of SCA, BSCA, maximum of 10 points. There are a few differences between and UCDA sensory analysis protocols, this article used the the protocols; BSCA begins with 30 points and UCDA has bootstrap method [9] to estimate the variation as a function one different score in sweetness regarding SCA protocol. of the number of tasters, which is a statistical technique of Nevertheless, the results are the same for all coffees; if >80 resampling with replacement used in several academic fields points, coffee is considered of specialty grade; if <80 points, [10]. coffee is considered below specialty quality. Usingtheframeworkpresentedabove,thestudyestimates For Conilon coffee, the quality was evaluated using the the optimal number of Q-Graders and R-Graders (the Q UCDA protocol, and the following attributes were obtained: Coffee System identifies quality coffees and brings them to total score, fragrance/aroma, flavor, balance, salinity/acidity, market through a credible and verifiable system; a common body, bitter/sweetness, and clean cup. standard for both Q Arabica (specialty grade) and Q Robusta In the case of samples of Arabica coffee and Conilon (Fine Robusta Grade) coffee has resulted in a universally coffee, which were adopted in the experiments, all presented shared language and standard top scoring lots) to be used in minimum scores of 80 points, ranging from 80 to 84 points, sensory tests that adopt SCA, BSCA, and UCDA methodolo- for both groups. gies. 2.3. Conduction of the Experiment with Q-Graders and R- Graders. The tests were performed in a sensorial laboratory 2. Materials and Methods ∘ under actinic artificial light, at 23 Candwithaircirculation. 2.1. Preparation of Samples. The studies with Q-Graders and The samples were arranged on two rectangular tables of R-Graders were conducted and elaborated in the Laboratory 110 cm in height, 70 cm in width, and 1.35 meters of length. Journal of Food Quality 3

For the accomplishment of the study, a blank experiment The results related to Figure 3 show that 5 tasters are was conducted, composed of 10 tasters who evaluated 20 needed to evaluate fragrance (a), flavor (b), and acidity (c) samples of Arabica coffee, with a minimum grade ≤ 80 points, and 6 tasters for sweetness (d) of Conilon coffee. considered specialty coffee according to SCA, BSCA, and Figure 4 shows the results regarding the amount of tasters UCDA sensory protocols. necessary to evaluate the gustatory quality of Conilon coffee Thetableofthe20samplesofConiloncoffeewascarried and, specifically, that 5 tasters are required to evaluate the out separately from Arabica coffee using the same 10 tasters attributes aftertaste (a), balance (c), and total score (d) and who evaluated the Arabica coffee samples. In the case of 6 tasters to evaluate clean cup (b). Conilon coffee, the cut-off score of >80 points was used to selectthe20samplesforcupping. 4. Discussion All the tasters used in the studies have certifications, either Q-Grader, R-Grader, or COB, as well as experience in These results confirm the work done by Bhumiratana etal. performing sensorial analyses with the following protocols: [16] and Di Donfrancesco et al. [2] who used six trained SCA, BSCA, and UCDA. testers to perform sensorial analysis of coffee. Further, these results are in parallel with Cook et al. [17], who adopted the 2.4. Ideal Number of Tasters. For the grouping of the pair use of 15 tasters initially and, after more careful selection numbers of tasters and their respective coefficients of vari- based on sensorial sensitivity, used 6 tasters for Arabica ation [𝑋, CV(𝑋)], we used the bootstrap method, where 1000 coffee. sample simulations were performed with 1, 2, 3, 4, 5, 6, 7, 8, 9, Other authors used a number of tasters inferior to and 10 tasters [13]. those found and recommended in this study. The works In order to determine the optimal size of the tasting room, of Bosselmann et al. [18], who adopted 3 tasters, used the the linear regression method of plateau response was used SCA methodology to perform the coffee quality analysis, [14]. The optimal number of tasters was proposed when the AlvaradoandLinnemann[3]used1tasterandajudgewith linear model becomes a plateau: 12 consumers trained to conduct their study, and Pereira et al. [19]adoptedtheuseof3tasters. {𝛽 +𝛽 𝑋 +𝜀 𝑋 ≤𝑋 0 1 𝑖 𝑖 se 𝑖 0 In the study by Ribeiro et al. [20], 11 experienced testers 𝑌𝑖 = { (1) 𝑃+𝜀 𝑋 >𝑋. were used to perform the sensorial analysis of coffee. { 𝑖 se 𝑖 0 However, many authors do not indicate if the tasters had 𝑌 𝛽 Here 𝑖 istheresponsevariable, 0 is the linear coefficient any type of international standardization in the materials and 𝛽 of the linear model of the segment before the plateau, 1 is methods, indicating once again the lack of standardization 𝜀 the angular coefficient of this same segment, 𝑖 is the error of the process of sensorial analysis of coffee. In the work of associated with the 𝑖th observation, 𝑃 is the plateau, and 𝑋0 is Pereira et al. [19] and Evangelista et al. [21], the authors do the point of attachment of the two segments. 𝑃 and 𝑋0 should not even identify the number of tasters used in their studies be estimated. to perform the sensorial analysis of coffee. For the statistical analysis, the free software R was used to Relevant considerations proposed by Chambers et al. [22] perform the bootstrap simulations and to obtain the statistics emphasize that in addition to the minimum number of tasters of the method of reaching the optimal number of tasters [15]. it is necessary to study the consistency of who is carrying out the analysis. For the authors, the three-member panel, 3. Results trained and experienced, had smaller residual error equal to one square of the semitrained panel for a sensory analysis Tables 1 and 2 show the results obtained from the 1000 sample study of chicken, turkey, and other birds, indicating that, simulations of the gustatory characteristics for Arabica and in addition to number, consistency should be respected and Conilon coffees, respectively, using the bootstrap method, widely observed. This indicates and reinforces the need to with 1,2,...,10 tasters and their respective coefficients of use professionals such as Q-Graders and R-Graders, since variation. these professionals are previously trained to perform such According to the results presented in Tables 1 and 2, it activities. This fact is also verified by Chambers et al. [23], is evident that the coefficient of variation as a function of because, for the authors, the training time contributes to the the number of Q-Graders and R-Graders decreases up to a levelofaccuracyoftheevaluator,reinforcingtheneedfor certain point, and from there the increase of the number of evaluators’ training. tasters for the sensorial analysis does not help to increase Thus it is evident that standardized methodologies can accuracy. give greater robustness to research and academic works that Figure 1 shows that 6 tasters are required to evaluate total use this technique. Many studies use Q-Graders, R-Graders, score (a) and fragrance (b) and 5 for flavor (c) and balance (d) and experts in sensory analysis and often do not provide of Arabica coffee using the linear regression method plateau. certification of the specialists. It is necessary to demand more In the graphs shown in Figure 2, it is estimated that veracity of sensory analysis of coffee, which is impossible 5 tasters are needed to evaluate the acidity (a), body (b), without consistency in the number of testers. sweetness (c), and aftertaste (d) characteristics of Arabica Nebesny and Budryn [24] have observed that there are coffee, by the same regression method. differences between the perceptions of women and men 4 Journal of Food Quality

Table 1: Grouping of the different numbers of Q-Graders and their respective coefficients of variation of the sensorial characteristics ofthe Arabica coffee samples with the SCA and BSCA protocols. Coefficient of variation (%) Number of Q-Graders Total note Fragrance Flavor Balance Acidity Body Sweetness Aftertaste 1 2.19 9.86 5.04 3.34 4.40 3.03 12.98 4.43 2 1.41 6.51 3.52 2.33 3.20 2.13 9.58 3.15 3 1.09 5.07 2.89 1.93 2.67 1.72 7.70 2.55 4 0.89 4.07 2.47 1.65 2.30 1.52 6.55 2.12 5 0.75 3.31 2.28 1.46 2.06 1.35 5.86 2.01 6 0.60 2.82 2.09 1.36 1.90 1.16 5.29 1.84 7 0.53 2.37 1.81 1.21 1.69 1.14 4.94 1.69 8 0.41 1.92 1.79 1.19 1.52 1.00 4.85 1.48 9 0.32 1.43 1.65 1.10 1.58 0.98 4.36 1.49 10 0.32 0.98 1.57 1.06 1.49 0.93 4.21 1.40

Table 2: Grouping of the different numbers of R-Graders and their respective coefficients of variation of the sensorial characteristics of Conilon coffee samples with the UCDA protocol. Coefficient of variation (%) Number of R-Graders Total note Fragrance Flavor Balance Salt/acid Body Bitter/sweet Clean cup 1 3.18 5.99 7.73 3.32 7.74 6.17 6.18 3.87 2 2.25 4.46 5.59 2.53 5.95 4.39 4.77 3.05 3 1.86 3.30 4.68 1.90 4.52 3.58 3.82 2.40 4 1.57 2.87 4.02 1.74 3.97 3.04 3.36 2.13 5 1.39 2.61 3.47 1.51 3.69 2.79 2.86 1.79 6 1.28 2.44 3.23 1.38 3.29 2.51 2.63 1.62 7 1.16 2.34 2.99 1.26 3.03 2.33 2.48 1.53 8 1.10 2.18 2.78 1.24 2.94 2.15 2.47 1.45 9 1.00 1.97 2.69 1.13 2.78 2.01 2.24 1.36 10 0.98 1.84 2.57 1.10 2.65 2.00 2.12 1.31

2,40 10,00 2,20 9,00 2,00 8,00 1,80 Y=2,2877 −0,340X ３？ X ≤ 5,4395 7,00 Y=10.4255 − 1.553X ３？ X ≤ 5,4854 Y=0, 1,60 4363 ３？ X > 5,4395 ≈6Q-Graders Y=1,9040 ３？ X > 5,4854 ≈6Q-Graders 2 ∗ 6,00 2 ∗ 1,40 R =0,8797 R =0,9035 5,00 1,20 1,00 4,00 0,80 3,00

CV (%) of fragrance CV (%) of 2,00 CV (%) of global note CV (%) of 0,60 0,40 1,00 0,20 0,00 1 2345678910 1 2345678910 Number of Q-Graders Number of Q-Graders (a) (b) 5,50 3,50 5,00 3,00 4,50 Y = 5,5645 − 0,8333X ３？ X ≤ 4,4401 Y=3,67473 − 0,544794X ３？ X ≤ 4,4887 4,00 Y = 1,8644 ３？ X > 4,4401 ≈ 5 Q-Graders Y=1,2293 ３？ X > 4,4887 ≈5Q-Graders 2 ∗ 2,50 2 ∗ R = 0,91617 R = 0,9114 3,50 3,00 2,00 CV (%) of flavor CV (%) of

2,50 balance CV (%) of 1,50 2,00 1,50 1,00 1 2345678910 1 2345678910 Number of Q-Graders Number of Q-Graders (c) (d) Figure 1: Relation between the coefficient of variation and number of tasters in Arabica coffee for global note (a), fragrance (b), flavor(c), ∗ and balance (d). Significant at 5%, by the 𝐹 test. Journal of Food Quality 5

4,50 3,50

4,00 3,00 3,50 Y=4,84472 − 0.681393X ３？ X ≤ 4.6074 2,50 Y=3,33511 − 0,494351X ３？ X ≤ 4,5357 Y=1,7053 ３？ X > 4,6074 ≈5Q-Graders Y=1,0929 ３？ X > 4,5357 ≈5Q-Graders 3,00 2 ∗ R2 = ∗ R = 0,9257 2,00 0,9071 2,50 1,50

2,00 body CV (%) of CV (%) of acidity CV (%) of 1,50 1,00 1,00 0,50 1 2345678910 1 2345678910 Number of Q-Graders Number of Q-Graders (a) (b) 13,00 4,50 12,00 4,00 11,00 3,50 10,00 Y=14,4980 − 2,11822X ３？ X ≤ 4.5229 Y=4.94244 − 0.751606X ３？ X ≤ 4.3776 Y=4.9175 ３？ X > 4.5229 ≈5Q-Graders Y=1.6522 ３？ X > 4.3776 ≈5Q-Graders 9,00 2 ∗ 3,00 2 ∗ R = 0.9457 R = 0.9353 8,00 2,50 7,00 2,00

6,00 aftertaste CV (%) of CV (%) of sweetness CV (%) of 5,00 1,50 4,00 1,00 1 2345678910 1 2345678910 Number of Q-Graders Number of Q-Graders (c) (d) Figure 2: Relationship between the coefficient of variation and number of tasters in Arabica coffee for acidity (a), body (b), sweetness (c), ∗ and aftertaste (d). Significant at 5%, by the 𝐹 test.

7,00 8,00

6,00 7,00

5,00 Y=6,78508 − 1,05214X ３？ X ≤ 4,3279 6,00 Y=8,51268 − 1,20214X ３？ X ≤ 4,6214 Y=2,2316 ３？ X > 4,3279 ≈5R-Graders Y=2,9571 ３？ X > 4,6214 ≈5R-Graders 2 ∗ 2 ∗ 4,00 R = 0,9480 5,00 R = 0,9244

3,00 4,00 CV (%) of flavor CV (%) of

CV (%) of fragrance CV (%) of 2,00 3,00

1,00 2,00 1 2345678910 1 2345678910 Number of R-Graders Number of R-Graders (a) (b) 8,00 7,00

7,00 6,00

6,00 Y=8,73029 − 1,27468X ３？ X ≤ 4,4466 Y=6,60831 − 0,804045X ３？ X ≤ 5,2493 Y=3,0623 ３？ X > 4,4466 ≈5R-Graders 5,00 Y=2,3876 ３？ X > 5,2493 ≈6R-Graders 2 ∗ 2 ∗∗ 5,00 R = 0,9301 R = 0,9402 4,00 4,00 CV (%) of acidity CV (%) of 3,00 sweetness CV (%) of 3,00

2,00 2,00 1 2345678910 1 2345678910 Number of R-Graders Number of R-Graders (c) (d) Figure 3: Relation between the coefficient of variation and number ofters tas in Robusta/Conilon coffee for fragrance (a), flavor (b), acidity ∗ ∗∗ (c), and sweetness (d). Significant at 5%; significant at 1%, by the 𝐹 test. 6 Journal of Food Quality

7,00 4,00

6,00

Y=6,84967 − 1,02144X ３？ X ≤ 4,4550 3,00 Y=4,1738 − 0,508146X ３？ X ≤ 5,3540 5,00 Y=2,2992 ３？ X > 4,4550 ≈5R-Graders Y=1,4532 ３？ X > 5,3540 ≈6R-Graders 2 ∗ 2 ∗∗ R = 0,9277 R = 0,9479 4,00 2,00 CV (%) of aftertaste CV (%) of 3,00 cup clean CV (%) of

2,00 1,00 1 2345678910 1 2345678910 Number of R-Graders Number of R-Graders (a) (b) 4,00 4,00

3,00 3,00 Y=3,71378 − 0,536227X ３？ X ≤ 4,5555 Y=3,52277 − 0,522871X ３？ X ≤ 4,5316 Y=1,2710 ３？ X > 4,5555 ≈5R-Graders Y=1,1534 ３？ X > 4,5316 ≈5R-Graders R2 = ∗ R2 = ∗ 0,9343 2,00 0,926649

2,00 1,00 CV (%) of balance CV (%) of CV (%) of global note CV (%) of

1,00 0,00 1 2345678910 1 2345678910 Number of R-Graders Number of R-Graders (c) (d) Figure 4: Relation between the coefficient of variation and number of tasters in Robusta/Conilon coffee for aftertaste (a), clean cup(b), ∗ ∗∗ balance (c), and global note (d). Significant at 5%; significant at 1%, by the 𝐹 test.

duringthesensorialanalysisofcoffee(inthearomaevalu- However, this approach is limited to the data of this study. In ation question). In the same line, Cook et al. [17] verified the case of limitations on the availability of Q-Graders or R- that differences in gender and age, as well as psychological Graders, the simulation and regression methods with plateau factors, have been attributed to the different perceptions of can be a solution model. Based on this the conclusions are as the sensorial analysis. As Pereira et al. [8] have described, follows. the noise factor has been pointed out as one of the greatest It is necessary to use 6 or more Q-Graders to perform the villains of the consistency of sensorial analysis with the use of sensory analysis with the SCA and BSCA protocol in scientific Q-Graders. studies and in routine taste tests for marketing purposes. It is plausible that judges’ perceptions may vary according For the UCDA protocol, the use of 6 or more R-Graders to these criteria, but this method is more standardized in the process of classification. for the sensory analysis of Conilon coffee is recommended, in In this way, it is possible to express from the results shown research and in sensory analysis for commercialization. that the number of 5 to 6 Q-Graders and/or R-Graders would Further studies should be developed about the accuracy be sufficient to ensure accuracy of the results of the sensory and consistency of Q-Graders and R-Graders. In addition, analysis (cup tests) and that the gains with the same would it is necessary to improve and approximate the scientific not be significant with the use of more testers for the decision methodsofvalidationofthelevelofaccuracyofthese making. As such, the results presented in Figures 1, 2, 3, and professionals, so that institutions that offer such courses can 4 show that the coefficient of variation decreases with more raise the level of precision of these professionals through less Q-Graders and/or R-Graders in sensory analysis. Moreover, subjective techniques. the optimum number of tasters occurred when the linear model becomes a plateau, and the linear regression method Conflicts of Interest of plateau response is a robust tool to quantify and validate the total need of tasters in sensory analysis. The authors declare that there are no conflicts of interest regarding the publication of this paper. 5. Conclusions Acknowledgments The modeling applied in this study allows concluding that according to the data tested, it is possible to recommend The authors acknowledge the Federal Institute of Esp´ırito the minimum number of evaluators, for these conditions. Santo for supporting this research and also the translation Journal of Food Quality 7

and review of this article, as well as the Q-Graders’ and R- Foundation for Statistical Computing, Vienna, Austria, 2016, Graders’ participation, who dedicated themselves to the real- https://www.R-project.org. ization of this study. They also acknowledge National Council [16] N. Bhumiratana, K. Adhikari, and E. Chambers, “Evolution of for Scientific and Technological Development (469058/2014- sensory aroma attributes from coffee beans to brewed coffee,” 5), CNPq, the Secretariat of Professional and Technological LWT- Food Science and Technology,vol.44,no.10,pp.2185– Education of the Ministry of Education, SETEC, for the 2192, 2011. availability of resources for research, and the Credit Unions [17]D.J.Cook,T.A.Hollowood,R.S.T.Linforth,andA.J.Taylor, “Correlating instrumental measurements of texture and flavour System in Brazil, SICOOB, for the support and funding of release with human perception,” International Journal of Food research. Science & Technology,vol.40,no.6,pp.631–641,2005. [18] A. S. Bosselmann, K. Dons, T. Oberthur, C. S. Olsen, A. References Ræbild, and H. Usma, “The influence of shade trees on coffee quality in small holder coffee agroforestry systems in Southern [1] A. M. Feria-Morales, “Examining the case of green coffee Colombia,” Agriculture, Ecosystems & Environment,vol.129,no. to illustrate the limitations of grading systems/expert tasters 1-3, pp. 253–260, 2009. in sensory evaluation for quality control,” Food Quality and [19] M. C. Pereira, S. M. Chalfoun, G. R. de Carvalho, and T. V. Preference,vol.13,no.6,pp.355–367,2002. Savian, “Multivariate analysis of sensory characteristics of coffee [2] B. Di Donfrancesco, N. Gutierrez Guzman, and E. Chambers, grains (Coffea arabica L.) in the region of upper Parana´ıba,” Acta “Comparison of results from cupping and descriptive sensory Scientiarum - Agronomy,vol.32,no.4,pp.635–641,2010. analysis of colombian brewed coffee,” Journal of Sensory Studies, [20] V. S. Ribeiro, A. E. Leitao,˜ J. C. Ramalho, and F. C. Lidon, vol. 29, no. 4, pp. 301–311, 2014. “Chemical characterization and antioxidant properties of a new [3] R. A. Alvarado and A. R. Linnemann, “The predictive value of a coffee blend with cocoa, coffee silverskin and green coffee small consumer panel for coffee-cupper judgment,” British Food minimally processed,” Food Research International,vol.61,pp. Journal,vol.112,no.9,pp.1023–1032,2010. 39–47, 2014. [4] R. N. Dal Molin, M. Andreotti, A. R. Reis, E. Furlani Junior, G. [21] S. R. Evangelista, C. F. Silva, M. G. P. D. C. Miguel et al., C. Braga, and M. B. Scholz, “Caracterizaçao˜ f´ısica e sensorial “Improvement of coffee beverage quality by using selected do cafe´ produzido nas condiçoes˜ topoclimaticas´ de Jesuitas, yeasts strains during the fermentation in dry process,” Food Parana,”´ Acta Scientiarum. Agronomy,vol.30,no.3,pp.353– Research International,vol.61,pp.183–195,2014. 358, 2008. [22]E.Chambers,J.A.Bowers,andA.D.Dayton,“Statistical [5]N.H.DzungandL.Dzuan,“Theroleofsensoryevaluationin Designs and Panel Training/Experience for Sensory Analysis,” food quality control, food research and development: a case of Journal of Food Science,vol.46,no.6,pp.1902–1906,1981. coffee study,” 2010, http://www4.hcmut.edu.vn/∼dzung/Senso- [23] D. H. Chambers, A.-M. A. Allison, and E. Chambers IV,“Train- ryrole.pdf. ing effects on performance of descriptive panelists,” Journal of [6] H. L. Meiselman, “Critical evaluation of sensory techniques,” Sensory Studies,vol.19,no.6,pp.486–499,2004. Food Quality and Preference, vol. 4, no. 1-2, pp. 33–40, 1993. [24] E. Nebesny and G. Budryn, “Evaluation of sensory attributes [7] C. F. Ross, “Sensory science at the human-machine interface,” of coffee brews from robusta coffee roasted under different Trends in Food Science & Technology,vol.20,no.2,pp.63–72, conditions,” European Food Research and Technology,vol.224, 2009. no. 2, pp. 159–165, 2006. [8] L.L.Pereira,W.S.Cardoso,R.C.Guarçoni,A.F.A.daFonseca, T.R.Moreira,andC.S.T.Caten,“Theconsistencyinthesensory analysis of coffees using Q-graders,” European Food Research and Technology,vol.243,no.9,pp.1545–1554,2017. [9]B.EfronandR.J.Tibshirani,An Introduction to the Bootstrap, Chapman and Hall-CRC, Washington, DC, USA, 1993. [10] R. L. Bastos and R. Lemos, Proposiçao˜ de testes bootstrap para o ´ındice de qualidade sensorial [Dissertaçao˜ de Mestrado], Universidade Federal de Lavras, UFLA, 2013. [11] SCA. Specialty Coffee Associationn., “Protocols,” 2013, http:// www.SCA.org/PDF/resources/cupping-protocols.pdf. [12] BSCA, Brazil Speciality Coffee Association, Varginha, Brasil, http://bsca.com.br. [13]F.d.Leonardo,W.E.Pereira,S.d.Silva,R.d.Araujo,´ and R. M. Mendonça, “Tamanho otimo´ da parcela experimental de abacaxizeiro Vitoria,”´ Revista Brasileira de Fruticultura,vol.36, no. 4, pp. 909–916, 2014. [14] P. F. Parnaiba, D. F. Ferreira, and A. R. Morais, “Tamanho otimo´ de parcelas experimentais: proposiçaoo˜ de metodos´ de estimac ¸ao˜ ,” Revista Brasileira de Biometria,vol.27,pp.255–268, 2009. [15] R Foundation for Statistical Computing, RCoreTeam.R: A Language And Environment for Statistical Computing,R Hindawi Journal of Food Quality Volume 2018, Article ID 6408571, 9 pages https://doi.org/10.1155/2018/6408571

Research Article Influence of Solar Radiation and Wet Processing on the Final Quality of Arabica Coffee

Lucas Louzada Pereira ,1 Rogério Carvalho Guarçoni,2 Wilton Soares Cardoso ,3 Renato Côrrea Taques,2 Taís Rizzo Moreira ,4 Samuel Ferreira da Silva,4 and Carla Schwengber ten Caten1

∘ 1 Federal University of Rio Grande do Sul, Av. Osvaldo Aranha, n 99/5 Andar, Bom Fim, 90035-190 Porto Alegre, RS, Brazil 2Capixaba Institute for Technical Assistance, Research and Rural Extension/Department of Statistics, Rua Afonso Sarlo, 160 Bento Ferreira, 29052-010 Vitoria,´ ES, Brazil 3Federal Institute of Esp´ırito Santo/Food Science and Technology Program, Rua Elizabeth Minete Perim, S/N. Bairro SaoRafael.29375-000VendaNovadoImigrante,ES,Brazil˜ 4Federal University of Esp´ırito Santo/Center of Agricultural Sciences and Engineers, Av. Gov. Lindemberg, 316 Centro, 29550-000 Jeronimoˆ Monteiro, ES, Brazil

Correspondence should be addressed to Lucas Louzada Pereira; [email protected]

Received 17 September 2017; Revised 5 December 2017; Accepted 10 January 2018; Published 11 February 2018

AcademicEditor:SilviaC.C.Botelho

The coffee growing in the state of´ırito Esp Santo has some peculiarities that differ from the other regions producing Arabica coffee in Brazil because it has a diversity of edaphoclimatic conditions that influence the final quality of the bean. This study aimed to demonstrate and quantify the effect of solar radiation and of different forms of wet process on the final quality of Arabica coffee in crops located in the altitude range of 950 meters, in order to understand what would be the best wet processing methods for the coffee cultivated to the East (sun-grown) and coffee cultivated to the South-Southeast (shade-grown). The results indicate that shading has a significant effect on the final quality of the Arabica coffee, as well as the type of wet process used to process thebeans after harvest. Therefore, there is a need to study in depth the factors related to the processing, edaphoclimatic, and relief conditions inherent to mountain coffee cultivation.

1. Introduction developed, aiming at better adaptation of the crop to the new global climate scenario. Although shading and altitude are Arabica coffee in Brazil is usually produced in full sun and in empirically known to have beneficial effects on coffee quality, areas that vary from 600 to 1400 meters of altitude, thus cre- only a few scientific studies have documented these effects ating strata and different sensory perceptions associated with [4]. Nevertheless, these kinds of studies need to consider the the flavors that each coffee can have due to edaphoclimatic level of solar radiation as well as the average temperature. and relief conditions and the amount of solar radiation that it It is known that, normally, coffee of colder regions receives. receives higher grades than samples from warmer regions, According to Bosselmann et al. [1] there is a great regarding taste, aroma, sweetness, and body [5]. This factor controversy about the production of shaded and sun-grown is associated with the fact that high temperatures prevent the coffee. On the other hand, for Somporn et al. [2], little has translocation of chemical compounds to fruits [1, 6, 7]. been discussed about the effect of shading on the final quality Another line of research has argued that in addition to the of the Arabica coffee. For Pinto Neto et al. [3], the observation factors mentioned above, it is necessary to study the effects of ofclimatechangeontheplanetcausesnewtechniquestobe fermentation in postharvest processing. The works of Silva et 2 Journal of Food Quality al.[8],Silvaetal.[9],andPereiraetal.[10]havediscussedthe The fresh water used in the processing of the coffees ∘ significant effect of yeasts during the fermentation process, in both experiments is in accordance with CONAMA n andthoseofRibeiroetal.[11],Evangelistaetal.[12],Masoud 357/2005 Resolution, which deals with the classification of and Jespersen [13], and finally De Bruyn et al. [14] discussed water bodies [17]. the effect of bacteria. These microorganisms were detected in Finally, both experiments were conducted in the same wet fermentation processes of the Arabica coffee, generating experimental design of randomized blocks, with the same significant impacts on coffee quality. Gonzalez-Rios et al. [15] number of replicates (05) and with the same treatments. found that the removal of mucilage through degradation in Experiment 02, East (sun-grown), received an average of water provided coffees with more fruity, floral, and caramel 22.3 more minutes of solar radiation per day than the 01, attributes and characteristics, while the removal of mucilage South-Southeast (shade-grown), representing 255.84 more 2 2 provided drier, more neutral beverages. Wh/m per day, corresponding to 93.381,6 Wh/m more per Thus, testing wet processing methods with induced fer- year. mentation, associated with the solar radiation index that the The South-Southeast and East experiments were con- crop receives, constitutes an innovative action. Therefore, ducted in a complete randomized block designed with five the hypothesis of this study is as follows: can the incidence replicates and four treatments, one with starter culture for of solar radiation, associated with different forms of wet fermentation of Arabic coffee, Yeast Fermentation (Saccha- processing with starter cultures, affect the quality of the romyces cerevisiae sp.), and the others with dry fermentation coffee? (Fully Washed) and with water (Washed) and pulped without This study had the objective of evaluating the effect of fermentation (Semidry). solar radiation, associated with four different forms of wet processing in the sensory quality of Arabic coffee. 2.1. Raw Materials: Wet Processing. The raw materials used in the formulation of the must were coffee pulp, coffee husk, 2. Materials and Methods water, and yeast (Saccharomyces cerevisiae sp.). Ten kilos of coffee were harvested per experimental plot The experiments were conducted on a property located at 950 in both experiments, presenting 85% of ripe fruits. After meters of altitude, as observed in Figure 1, using the variety harvest, the fruits were processed according to the treatment. Catua´ı red 44. Table 1 indicates the geographic coordinates, the average duration of annual radiation, the average daily 2.2. Preparation of Musts. Of the four proposed treatments, solar radiation for the year, and total solar radiation per year. one was prepared from the must, according to the process The climate of the region of Experiment 1 is characterized of patent BR1020160040531 (African drying beds), with yeast as hot and humid, with annual rainfall of 1200 to 1300 mm and ∘ culture (Saccharomyces cerevisiae sp.) and coffee husk. The average annual temperature of 19 C. The climate of the region four treatments followed the following methods. of Experiment 2 is also characterized as hot and humid, with an annual rainfall of 1200 to 1300 mm, and average annual Treatment 01.Itisdryfermentationmust(FullyWashed,FW), ∘ temperature of the 19 C. 10 kg of peeled cherry coffee (pulp), and 5 kg of husk, without The soil samples from both experiments which were taken adding water to the process. from the depth of 0–20 cm before the implantation of the experiment were analyzed. Treatment 02.Itisfermentationmustwithwater(Washed, W), 10 kg of peeled cherry coffee (pulp), 5 kg of husk, and5 The experiments and the management of the fertilization liters of water. were carried out according to the results of the soil analysis, according to the Liming and Fertilization Manual for the State Treatment 03. It is fermentation must with yeast starter of Esp´ırito Santo-5th approximation [16]. Fertilization was culture, Saccharomyces cerevisiae sp., (Yeast Fermentation, carried out in three parts from October to March. The soil YF), 1% of the must (p/v), 10 kg of peeled cherry coffee (pulp), correction of both experiments was carried out according to 5 kg of husk, 100 grams of yeast, and 5 liters of water. the Liming and Fertilization Manual for the State of Esp´ırito Santo-5th approximation [16] according to the results of soil Patent process (African drying beds) was deposited at the analysis. National Institute of Industrial Property. Liming was performed in June for both experiments since the base saturation for the coffee crop should be 푉 =60% Treatment 04.Itispulpedcoffeewithoutmucilagewithdrawal [16]. The phytosanitary control was conducted in October in (Semidry, SD) and without any addition of microorganisms. ∘ a preventive way as typical in the region. Musts 02 and 03 received water at 38 Candremained For the experimental control of the experiments it was immersed in plastic fermentation tanks in the laboratory for observed that both are located at 950-meter altitude; the soil 36 hours. The temperature of the fermentation room was ∘ type of the two experimental areas is dystrophic Red-Dark stabilized at 21 C. Podzolic; the soil fertility conditions were corrected through Treatments 01, 02, and 03 remained in sealed tanks for fertilization and liming according to Prezotti et al. [16] in the 36 hours inside the fermentation room and, after this period, two experiments; the climatic variations (temperature and they were washed and taken to dry in African drying beds, rainfall) are similar in both areas, since they are close. such as in the farms. Journal of Food Quality 3

41∘100７41∘00７40∘500７

∘   20 10 30 ３ 2

Brazil Espírito Santo

20∘210３ 1

Digital elevation model Experiment High: 2857

Low: 0

∘ Experiment Annual rainfall (mm) Average annual temperature ( C) Soil types 1 1200–1300 19 Red-Dark Podzolic 2 1200–1300 19 Red-Dark Podzolic

N (km) Geographic coordinates system 0 50 100 Datum: WGS 84 ZONA 24S

Figure 1: Location of the study area.

Table 1: Geographical coordinates and incidence of solar radiation for the two experiments.

Solar radiation, SR, daily Average annual radiation SR total year Experiment Latitude Longitude average for the year 2 duration (hours) 2 (WH/m ) (Wh/m ) 01 −20.3263 −41.1069 10.3673 4285.12 1564068.8 02 −20.1667 −41.0375 10.7385 4540.96 1657450.4 Source: [18]; Wh: Watt-hour.

Treatment 04 was pulped and taken to dry African drying evidence protocol of the Specialty Coffee Association of beds. The process of fermentation occurred naturally, without America, SCAA. All tasters tested the 20 plots of each any addition of microorganisms. experiment. The SCAA protocol determines the overall quality of the coffee through an additive sum of 10 sensory 2.3. Sensory Analyses. The sensory analyses were performed parameters consisting of fragrance, flavor, aftertaste, acidity, by 06 Q-Graders, according to the method proposed by body,uniformity,cleancup,sweetness,balance,andoverall. Pereira et al. [19]. The authors determined the number with simulation of the Bootstrap method. 2.4. Statistical Analysis. A joint analysis of the experiments TheevaluationswerecarriedoutintheLaboratory was performed. The results were compared with the Tukey of Research and Analysis in Coffee, LAPC, following the test at 5% probability, followed by analysis of the main 4 Journal of Food Quality

Table 2: Grades of the means of the sensory attributes for each treatment related to the crop located to the South-Southeast.

Pulped without fermentation Dry fermentation Fermentation with Attributes Yeast Fermentation (Semidry) (Fully Washed) water (Washed) Fragrance 7.94 7.99 8.29 8.01 Flavor 7. 72 7. 8 0 7. 9 9 7. 8 0 Aftertaste 7. 62 7. 5 6 7. 78 7. 6 6 Acidity 7. 8 1 7. 8 0 7. 9 8 7. 72 Body 7. 62 7. 6 1 7. 8 0 7. 6 5 Uniformity 9.78 9.79 9.96 9.88 Clean cup 9.85 9.86 9.92 9.93 Sweetness 9.40 9.44 9.62 9.49 Balance 7. 5 8 7. 5 8 7. 74 7. 6 3 Overall 7. 6 1 7. 6 6 7. 8 5 7. 6 4

Table 3: Grades of the means of the sensory attributes for each treatment related to the crop located to the East.

Attributes Peeled without Dry process Fermentation with Fermentation with yeast fermentation (Semidry) (Fully Washed) water (Washed) (Yeast Fermentation) Fragrance 7. 8 1 7. 8 5 7. 7 7 7. 78 Flavor 7. 6 1 7. 6 6 7. 6 5 7. 6 3 Aftertaste 7. 4 9 7. 4 5 7. 4 1 7. 4 3 Acidity 7. 6 4 7. 7 7 7. 6 5 7. 5 5 Body 7. 5 0 7. 5 1 7. 4 5 7. 5 7 Uniformity 9.85 9.85 9.78 9.81 Clean cup 9.88 9.95 9.57 9.89 Sweetness 9.42 9.42 9.51 9.50 Balance 7. 4 9 7. 5 3 7. 3 9 7. 52 Overall 7. 4 5 7. 3 6 7. 4 1 7. 52 components to group the treatments, using visual exams coffee quality. It is known that usually the coffee in the in graphical dispersions for the South-Southeast and East colder and shaded region (higher altitude) receives higher experiments, considering an accumulated variability above grades regarding flavor, aroma, sweetness, and body than 70%adequatetoperformtheanalysis.Forstatisticalanalysis, samples from warmer regions. For DaMatta [5], this factor theSAEGsoftwarewasused[11]. is associated with the fact that the high temperatures prevent the translocation of chemical compounds to the fruits. 3. Results and Discussion The results expressed in mean values for overall quality between the South-Southeast and East experiments were not The results of the sensory analyses are presented in Tables 2 statistically different from each other and did not confirm and3,whichshowthemeanvaluesofthesensoryattributes the same results found by Evangelista et al. [21] and Pereira by treatment. All average results place the coffees in the range et al. [22]. To these authors, the fermentation induced with of specialty coffees by the SCAA protocol. yeast culture promoted quality gains to the observed coffees. Table 4 presents the average results of the overall quality This indicates that further studies are necessary to understand for the four treatments in the two environments, South- the modifications that may occur during the fermentation Southeast and East. phase, providing greater clarification on the action of the The results indicated that for the South-Southeast envi- microorganisms, as well as the competition between microor- ronment the fermentation must with water was superior in ganisms that occurs during the fermentation phase with relation to pulped coffee without fermentation. However, it cultures starters. did not differ from the other treatments. It was verified in the experiment located to the East that For the experiment located in the South-Southeast, the the treatments did not differ among themselves at 5% of treatment with water presented higher overall quality than probability. the experiment located in the East. In this way, the results Table5showstheeigenvalueswithsimpleandcumulative corroborate the propositions of Joet¨ et al. [4] and Muschler percentages of the total variance of the main components of [20], indicating that the shading has exerted influence on the South-Southeast experiment. Journal of Food Quality 5

Table 4: Averages of the overall quality evaluated in four treatments and in two environments, Venda Nova do Imigrante, in 2015.

Global quality Treatment Average South-Southeast East abA aA a Dry fermentation (Fully Washed) 83.06 82.16 82.61 aA aB a Fermentation with water (Washed) 84.86 81.63 83.25 abA aA a Yeast Fermentation 83.5 0 82.24 82.87 bA aA a Pulped without fermentation (Semidry) 82.63 81.70 82.17 A B Average 83.51 81.93 Means followed by at least one same horizontal capital letter and at least one same lowercase vertical letter do not differ from one another by the Tukey test at 5% probability.

Table 5: Main components, their respective eigenvalues, and simple and accumulated percentages of the total variance of the South-Southeast experiment.

Main component Eigenvalues Simple percentage Accumulated percentage CP1 9.998819 99.98819 99.98819 CP2 0.0008436559 0.00844 99.99662

Table 6: Main components, their respective eigenvalues, and simple and accumulated percentages of the total variance of the East experiment.

Main Component Eigenvalues Simple percentage Accumulated percentage CP1 9.998897 99.98897 99.98897 CP2 0.001005009 0.01005 99.99902

0.05 0.04 0.03 YF 0.02 SD 0.01

CP 2 0.00 −0.01 −0.02 FW −0.03 −0.04 W −0.04 −0.03 −0.02 −0.01 0.00 0.01 0.02 CP1 Figure 2: Diagram of dispersion in relation to the first two main components of the treatments: dry fermentation must—Fully Washed (FW), fermentation must with water—Washed (W), fermentation must with yeast—Yeast Fermentation (YF), and pulped without fermentation—Semidry (SD), from the South-Southeast (shade-grown) experiment.

Figure 2 presents the dispersion of the treatments of the Table 6 shows the eigenvalues and simple and accumu- South-Southeast experiment, based on the respective coor- lated percentages of the total variance of the East experiment. dinates related to the first two main components, CP1 and Figure 3 shows the dispersion of the treatments of the East CP2. The treatments were treatment with dry fermentation experiment based on the respective coordinates relative to the must (Fully Washed, FW), fermentation must with starters first two main components CP1 and CP2. The treatments with yeast cultures (Yeast Fermentation, YF), and pulped coffee dry fermentation must (Fully Washed, FW), fermentation without fermentation (Semidry, SD) form one group and the must with starters yeast cultures (Yeast Fermentation, YF), treatment fermentation with water (Washed, W) in another andpulpedcoffeewithoutfermentation(Semidry,SD)form group. In addition, the two components absorbed 99.997% of a group and the treatment fermentation with water (Washed, the variation existing in the original features (Table 5). W) forms another group, and the two components absorbed The results expressed in Figure 2 indicate that for the 99.999% of the variation existing in the original features condition of the experiment located in the South-Southeast, (Table 6). that is, shaded terrains, the fermentation with water, the The analyzed data of the eastern crop show results similar method commonly adopted in Colombia, is grouped far from to those of the experiment located in the South-Southeast, in the other treatments. relationtothemaincomponents.However,inthetreatment 6 Journal of Food Quality

0.05 0.04 0.03 YF 0.02 SD 0.01

CP 2 0.00 −0.01 −0.02 FW −0.03 −0.04 W −0.04 −0.03 −0.02 −0.01 0.00 0.01 0.02 CP1 Figure 3: Diagram of dispersion in relation to the first two main components of the treatments: dry fermentation must—Fully Washed (FW),fermentationmustwithwater—Washed(W),fermentationmustwithstartersofyeast—YeastFermentation(YF),andpulpedwithout fermentation—Semidry (SD), from the East experiment (sun-grown).

with water, the overall quality was higher for the South- sugars present in the pulp, thereby creating metabolic routes Southeast environment when compared to the East, as shown and differentiated sensory patterns. For Velmourougane [25], in Table 4. This result can be explained by the action of it is important to study and understand the fermentation the microorganisms to the detriment of the microbiota that process to develop flavor and a high quality standard of coffee. is present in the two experiments. Evangelista et al. [21] It is evident that fermentation is a complex process, have argued that the formation of the microbiota, the room involving several factors, with the action of different microor- temperature, and the pH of the must may undergo changes ganisms that can act in the improvement, as in the loss, and modify the intensity and the way the microorganisms act of quality. Consequently, it can contribute or not to the during the fermentation process. The fruits of the coffee when deterioration of the product’s final taste, due to the kind of being processed allow the emergence of a spontaneous or wild actiontheycantakeonprocessing.Hencetheimportanceof fermentation. The sugars and pectins present in the mucilage knowing more about the action of the microbiota and about allow the growth of microorganisms, especially bacteria and fermentation processes during the production of specialty yeasts. coffees,inviewoftheopportunitytocreatemorestandard- Yet, the scientific literature contains virtually no studies ized processes that consider the described factors, aiming to on microbial interactions at a single cell level occurring on promote quality improvements to the final product, as well as solid and liquid surfaces, involving fermentation microor- food safety to the consumer. ganisms [23]. Developing and controlling processes in the production This fact may have influenced positively the fermentation ofspecialtycoffeeshaveprovedtobeacomplextaskand with water for the experiment located in the South-Southeast often without a consensus on the best postharvest pro- and influenced it negatively in the experiment located in the cessing method, which may vary from region to region. East, confirming the need for greater monitoring during the Velmourougane [25] reinforces the need to study processing fermentation stages of Arabica coffee [14]. techniques, looking for ways to improve the final quality of Precisely because of the lack of clarification about the Arabica coffee. medium of fermentation within the coffee fruit, this hypoth- Results presented in Figure 4 corroborate with the data esis proposes a new perspective for science, the monitoring presented in Table 4, showing the values of the most signifi- of the microbiota that is formed inside the coffee fruits and cant sensory attributes for the fermentation with water treat- how it develops after harvesting under different fermentation ment, in relation to the nonfermented (semidry), confirming conditions. the data of de Melo Pereira et al. [10] and Lee et al. [24]. According to the perspective of Lee et al. [24], the effects Figure 5 shows the dispersion of the sensory attributes of fermentation during the wet treatment on the aroma profile of the treatments related to the experiment in the East, ofcoffeearenotcompletelyelucidatedandareoftenneglected corroborating with the results obtained in Table 4 for the since the literature has argued that the main function of overall quality of the Arabica coffee. fermentation is the removal of the mucilage. For De Bruyn It is possible to understand that the washed fermentation et al. [14], further studies should be undertaken to strengthen canbeanenginemakerofnewroutesandnuancesforthe the understanding of the impact of the microbiota on coffee formation of the sensory features of the wet processed coffee, quality and provide robust data for the development of more aiming at the potentialization and optimization of the quality controlled fermentation processes. curve of Arabica coffee. The coffee fermentation occurs to solubilize polysaccha- On the difference between the fermentation methods, the rides, which are present in the coffee pulp, after removal of results of Somporn et al. [2] indicated that shading exerts the mucilage, facilitating the drying of the fruits. During the an influence on the formation of sugars, chlorogenic acids, fermentation, microorganisms act in the degradation of the and total phenolics. In addition to shading, DaMatta [5] Journal of Food Quality 7

Fragrance

Overall Uniformity

Balance Body

Aftertaste Sweetness

Acidity Flavor

Clean cup

Washed Yeast Fermentation Fully Washed Semidry

Figure 4: Average dispersion of attributes: fragrance, uniformity, clean cup, sweetness, flavor, acidity, body, aftertaste, balance, and overall, of the coffees located in the South-Southeast (shade-grown).

Fragrance Uniformity Overall

Balance Body

Aftertaste Sweetness

Acidity Flavor Clean cup

Washed Yeast Fermentation Fully Washed Semidry

Figure 5: Average dispersion of the attributes: fragrance, uniformity, clean cup, sweetness, flavor, acidity, body, aftertaste, balance, and overall, of the coffees situated to the East (sun-grown).

argues that fruits that are formed under conditions of higher characteristics, in relation to the experiment located in the temperatures mature prematurely, preventing the complete East, area with higher solar incidence. translocation of compounds responsible for the typical aroma and flavor of the coffee. 4. Conclusion Thus, the microbiota of the experiment located in the South-Southeast associated with the water-induced fermen- The overall quality of the coffees presented the most promis- tation method contributed to the modification of the sensory ing results for wet processing through water fermentation in 8 Journal of Food Quality relation to the nonfermentation method (Semidry) for the qualitedescaf´ es´ Arabica,” Juillet. Aout,ˆ Plantations, Recherche, experiment located in the South-Southeast region. Developpement´ ,vol.3,no.4,pp.272–283,1996. The lower incidence of solar radiation in the crop had [7] B. Bertrand, P. Vaast, E. Alpizar, H. Etienne, F. Davrieux, and a significant effect on the overall quality of the Arabica P. Charmetant, “Comparison of bean biochemical composition coffee, associated with wet processing and water fermentation and beverage quality of Arabica hybrids involving Sudanese- (Washed). Ethiopian origins with traditional varieties at various elevations These results demonstrate and reinforce the condition in Central America,” Tree Physiology,vol.26,no.9,pp.1239– of the environment; that is, the incidence of solar radia- 1248, 2006. tion can lead to changes in internal metabolites, creating [8] C.F.Silva,L.R.Batista,L.M.Abreu,E.S.Dias,andR.F.Schwan, “Succession of bacterial and fungal communities during natural a stress condition, and consequently different conditions coffee (Coffea arabica) fermentation,” Food Microbiology,vol. for the development of microorganisms. However, this new 25,no.8,pp.951–957,2008. hypothesisneedstobebetterclarified. [9]C.F.Silva,L.R.Batista,andR.F.Schwan,“Incidenceand Variationswithrespecttothewaterfermentationmethod distribution of filamentous fungi during fermentation, drying for the experiment located in the South-Southeast and East and storage of coffee (Coffea arabica L.) beans,” Brazilian regions may be related to the action of the microorganisms Journal of Microbiology,vol.39,no.3,pp.521–526,2008. present in the Arabica coffee and new research and studies [10] G. V. M. Pereira, V. T. Soccol, A. Pandey et al., “Isolation, are needed to deepen and to quantify the action of these selection and evaluation of yeasts for use in fermentation of microorganisms in loco. coffee beans by the wet process,” International Journal of Food Microbiology, vol. 188, pp. 60–66, 2014. Conflicts of Interest [11] B. B. Ribeiro, L. M. V. L. Mendonça,G.A.Assis,J.M.A.de Mendonça,M.R.Malta,andF.F.Montanari,“Evaluationofthe The authors declare that there are no conflicts of interest chemical and sensory characteristics of Coffea canephora Pierre regarding the publication of this paper. and Coffea Arabica L. blends,” Coffee Science,vol.9,no.2,pp. 178–186, 2014. [12] S. R. Evangelista, M. G. da Cruz Pedrozo Miguel, C. de Souza Acknowledgments Cordeiro,C.F.Silva,A.C.MarquesPinheiro,andR.F.Schwan, “Inoculation of starter cultures in a semi-dry coffee (Coffea The authors thank the Federal Institute of´ırito Esp Santo arabica) fermentation process,” Food Microbiology,vol.44,pp. for supporting this research and also for the translation 87–95, 2014. and review of this article. They also thank National Council [13] W. Masoud and L. Jespersen, “Pectin degrading enzymes in for Scientific and Technological Development (469058/2014- yeasts involved in fermentation of Coffea arabica in East Africa,” 5), CNPq, and Secretariat of Professional and Technological International Journal of Food Microbiology,vol.110,no.3,pp. Education of the Ministry of Education (SETEC) for the 291–296, 2006. availability of resources for research. [14] F. De Bruyn, S. J. Zhang, V. Pothakos et al., “Exploring the impacts of postharvest processing on the microbiota and metabolite profiles during green coffee bean production,” References Applied and Environmental Microbiology,vol.83,no.1,Article [1] A. S. Bosselmann, K. Dons, T. Oberthur, C. S. Olsen, A. ID e02398-16, 2017. Ræbild, and H. Usma, “The influence of shade trees on coffee [15] O. Gonzalez-Rios, M. L. Suarez-Quiroz, R. Boulanger et quality in small holder coffee agroforestry systems in Southern al., “Impact of “ecological” post-harvest processing on coffee Colombia,” Agriculture, Ecosystems & Environment,vol.129,no. aroma: II. Roasted coffee,” JournalofFoodCompositionand 1-3, pp. 253–260, 2009. Analysis,vol.20,no.3-4,pp.297–307,2007. [2]C.Somporn,A.Kamtuo,P.Theerakulpisut,andS.Siriamorn- [16]L.C.Prezotti,J.A.Gomes,G.G.Dadalto,andJ.A.Oliveira, pun, “Effect of shading on yield, sugar content, phenolic acids “Manual de recomendaçao˜ de calagem e adubaçao˜ para o and antioxidant property of coffee beans (Coffea Arabica L. cv. Estado do Esp´ırito Santo – 5a aproximaçao,”˜ Vitoria:´ SEEEA/ Catimor)harvestedfromnorth-easternThailand,”Journal of the INCAPER/CEDAGRO, article 305, 2007. Science of Food and Agriculture,vol.92,no.9,pp.1956–1963, [17] CONAMA – CONSELHO NACIONAL DE MEIO AMBIENTE 2012. (Brasil), “Resoluçaon357,de17demarc˜ ¸o de 2005,” Diario´ [3]J.N.PintoNeto,M.I.N.Alvarenga,M.P.Correa,andC.C. Oficial da Uniao˜ , pp. 58–63, 2015. Oliveira, “Efeito das variaveis´ ambientais na produçao˜ de cafe´ [18] GEOBASES. Sistema Integrado de Bases Geoespaciais do em um sistema agroflorestal,” Coffee Science,vol.9,no.2,pp. Estado do Esp´ırito Santo. 2016, http://www.geobases.es.gov.br/ 187–195, 2014. portal/. [4] T. Joet,¨ A. Laffargue, F. Descroix et al., “Influence of environ- [19] L. L. Pereira, R. C. Guarçoni, I. C. Pulini et al., “Tamanho otimo´ mental factors, wet processing and their interactions on the do numero´ de provadores de cafe´ com uso do protocolo SCAA,” biochemical composition of green Arabica coffee beans,” Food Simposio Internacional De Ingenier´ıa Industrial: Actualidad Y Chemistry,vol.118,no.3,pp.693–701,2010. Nuevas Tendencias,vol.9,2016. [5]F.M.DaMatta,“Exploringdroughttoleranceincoffee:A [20] R. G. Muschler, “Shade improves coffee quality in a sub-optimal physiological approach with some insights for plant breeding,” coffee-zone of Costa Rica,” Agroforestry Systems,vol.51,no.2, Brazilian Journal of Plant Physiology,vol.16,no.1,pp.1–6,2004. pp.131–139,2001. [6] B. Guyot, D. Gueule, J. C. Manez, J. J. Perriot, J. Giron, [21] S. R. Evangelista, C. F. Silva, M. G. P. D. C. Miguel et al., and L. Villain, “Influence de l’altitude et de l’ombrage sur la “Improvement of coffee beverage quality by using selected Journal of Food Quality 9

yeasts strains during the fermentation in dry process,” Food Research International,vol.61,pp.183–195,2014. [22] G. V. M. Pereira, E. Neto, V. T. Soccol, A. B. P. Medeiros, A. L. Woiciechowski, and C. R. Soccol, “Conducting starter culture- controlled fermentations of coffee beans during on-farm wet processing: Growth, metabolic analyses and sensorial effects,” Food Research International,vol.75,pp.348–356,2015. [23] R. F. Schwan and G. H. Fleet, Cocoa and Coffee Fermentation, Taylor & Francis Group, LLC. CRC Press, 2015. [24] L. W.Lee, M. W.Cheong, P.Curran, B. Yu, and S. Q. Liu, “Coffee fermentation and flavor - An intricate and delicate relationship,” Food Chemistry,vol.185,pp.182–191,2015. [25] K. Velmourougane, “Impact of natural fermentation on physicochemical, microbiological and cup quality characteristics of Arabica and Robusta coffee,” Proceedings of the National Academy of Sciences India Section B - Biological Sciences,vol.83, no. 2, pp. 233–239, 2013.